Chirality in Anesthesia II: Stereoselective Modulation of Ion Channel Function by Secondary Alcohol Enantiomers : Anesthesia & Analgesia

Secondary Logo

Journal Logo

Anesthetic Pharmacology: Research Report

Chirality in Anesthesia II: Stereoselective Modulation of Ion Channel Function by Secondary Alcohol Enantiomers

Brosnan, Robert DVM, PhD*; Gong, Diane PharmD; Cotten, Joseph MD, PhD; Keshavaprasad, Bharat MD§; Yost, C Spencer MD§; Eger, Edmond I II MD§; Sonner, James M. MD§

Author Information
Anesthesia & Analgesia 103(1):p 86-91, July 2006. | DOI: 10.1213/
  • Free

Chirality has been proposed as a means for distinguishing molecular targets that are relevant to anesthetic action from those that are not. In applying this approach, patterns of enantioselectivity in animals are matched with the pattern of enantioselectivity on molecular targets. For example, the similar pattern of enantioselectivity observed between loss of the righting reflex in rats and the function of γ-amino butyric acid type A (GABAA) receptors and baseline potassium channels for isoflurane has been cited in support of those channels as mediators of isoflurane's effect on the righting reflex and used to argue against a role for two other channels that did not display enantioselectivity (1).

The isomers of isoflurane are not commercially available, which restricts their use in confirming targets of anesthetic action by this approach. In a companion paper (2) to the present essay, we report that two experimental volatile anesthetics that are readily available, 2-butanol and 2-pentanol, are enantioselective with respect to the minimum alveolar concentration preventing movement in 50% of individuals (MAC) in rats. The differences in MAC for these compounds are similar to those reported for isoflurane, with a 17% change in MAC for the 2-butanol isomers and a 38% change in MAC for the 2-pentanol isomers, compared with a MAC difference reported as 17% in one study (3) and 53% in another study (4) for the isomers of isoflurane. Enantioselectivity for these inexpensive secondary alcohol isomers potentially enables a large number of studies of chirality in anesthesia.

However, the usefulness of chirality as a test of relevance of molecular targets to anesthesia relies, as does any diagnostic test, on the sensitivity and specificity of the test (5). These numbers are unknown but have been implicitly assumed to be high in studies of chirality in anesthesia. Today, the hypothesis that chirality is sufficiently sensitive and specific to be a test of relevance of molecular targets can be tested because molecular targets that are thought to have a high prior probability of being mediators of anesthetic action (i.e., ion channels thought to mediate MAC and other anesthetic actions) have been identified and the inexpensive chiral anesthetic alcohols are available in sufficient quantities to test these multiple molecular targets and to perform studies in large numbers of animals. If indeed chirality were a sensitive and specific test of relevance of anesthetic targets, then these molecular targets should display patterns of enantioselectivity similar to these seen in animals.

We applied the enantiomers of 2-butanol, 2-pentanol, and 2-hexanol at MAC concentrations to 3 ion channels. 2-Butanol and 2-pentanol are enantioselective with respect to MAC, whereas 2-hexanol is not (2). We hypothesized that if enantioselectivity were a good test of anesthetic relevance, enantioselectivity would be observed on these channels by 2-butanol and 2-pentanol but not by 2-hexanol.

The three anesthetic-sensitive ion channels selected are representative of ion channel families thought to mediate anesthetic effects generally, including MAC. They included the Twik-Related-spinal cord K+ (TRESK) channel, a GABAA receptor, and an N-methyl-d-aspartate (NMDA) receptor. The human TRESK channel is a 2-pore domain potassium leak channel formed from dimers of pore-forming units, each containing four transmembrane segments and two tandem P domains (the putative pore-forming region) (6). TRESK currents are enhanced by volatile anesthetics (7). GABAA receptors are heteropentameric chloride channels gated by γ-amino butyric acid (GABA). Functional channels can be formed by α and β subunits; physiologically, a γ subunit is often also present. Anesthetics enhance chloride currents in response to activation of receptors by GABA (8). NMDA receptors are selectively permeable to calcium. They are gated by l-glutamate and the coagonist glycine, which bind the NR2 and NR1 subunits, respectively, both of which are required for function. Volatile anesthetics inhibit NMDA receptor function (9–10).


The Animal Use and Care Committee at the University of California, San Francisco approved this project. Eggs from Xenopus laevis frogs were defolliculated using 0.2% collagenase A (Worthington Biochemical Corportation, Lakewood, NJ) for approximately 1 h and stored in modified Barth solution (88 mM NaCl, 1 mM KCl, 2.4 mM NaHCO3 2.4 mM, HEPES 20 mM, MgSO4 0.82 mM, Ca(NO3)2 0.33 mM, CaCl2 0.41 mM, sodium pyruvate 5 mM, gentamycin, penicillin, streptomycin, filtered, pH = 7.4). Stage V and VI oocytes were selected for nucleic acid injection and study. All experiments were conducted at room temperature.

The human TRESK gene was obtained from reverse transcription polymerase chain reaction of central nervous system tissue RNA and cloned into a pCDNA3.1-TOPO vector (7). Plasmids were linearized with Pme I, and RNA was synthesized using a T7 RNA polymerase from a commercially available kit (mMessage mMachine; Ambion; Austin, TX). Oocytes were injected with 1–4 ng of hTRESK cRNA in water and studied 1–3 days later. Water-injected oocytes served as controls.

Oocytes were placed in a 250 μL perfusion chamber and perfused with frog Ringer's solution (115 mM NaCl, 2.5 mM KCl, 1.8 mM CaCl2, 10 mM HEPES, filtered, pH = 7.4) at 2 mL/min. A standard two electrode voltage clamping technique using a GeneClamp 500B amplifier (Axon Instruments, Foster City, CA) maintained a –60 mV transmembrane potential in oocytes. Each minute a 1-s voltage pulse step protocol was performed from –100 to + 60 mV with 20-mV intervals and 1.5-s interpulse intervals (GeneClamp 500B; Axon Instruments, Union City, CA). Data were digitally acquired (Powerlab data acquisition hardware and Scope software; AD Instruments, Colorado Springs, CO). Currents measured during the +60 clamp were used for subsequent analysis. All hTRESK oocytes studied had peak currents between 1–10 μA at +60mV.

After a wash with frog Ringer's solution to ensure baseline drifts were <10% over 5 min, we separately delivered the R(−) alcohol, the S(+) alcohol, and the racemic alcohol. The test alcohols, each given separately at 2 mL/min for 2 min, were 16 mM 2-butanol (99% purity), 4 mM 2-pentanol (98% purity), and 2 mM 2-hexanol (99% purity). All were obtained from Sigma-Aldrich (St. Louis, MO). Recordings taken at the end of the 2-min delivery period were used for analysis. Finally, the oocyte was perfused again with frog Ringer's solution for 10 min. Washout of drug effects was considered acceptable if the current returned to within 10% of the pre-alcohol exposure current.

Clones of murine NR1 in a pCDNA3 vector and NR2A in a Bluescript vector were kindly provided by the R. A. Harris (Institute for Cellular and Molecular Biology, Waggoner Center, University of Texas, Austin) and C. S. Yost (Department of Anesthesia and Perioperative Care, UCSF) laboratories. Plasmids were linearized with ApaI and XhoI, respectively, and RNA was synthesized using T7 RNA from a commercially available kit (mMessage mMachine; Ambion; Austin, TX). Transcripts were mixed in a 1:1 ratio, and 1–8 ng of total RNA was injected into oocytes that were studied 1–2 days later.

A two-electrode voltage clamping apparatus identical to that described previously was used (but with Chart software, AD Instruments). During a –80 mV clamp, oocytes were perfused with a barium frog Ringer's solution (115 mM NaCl, 2.5 mM KCl, 1.8 mM BaCl2, 10 mM HEPES, 0.1 mM EGTA, filtered, pH = 7.4). Using a programmable, automated perfusion system (Valve Bank, AutoMate Scientific, San Francisco, CA), the Ringer's solution was switched to one containing agonist (barium frog Ringer's solution plus 0.1 mM glutamate and 0.01 mM glycine) 3 times at 6-min intervals to verify constancy of the baseline response. Six minutes after this last exposure, oocytes were perfused with the test alcohol solution in barium frog Ringer's for 1 min and 40 s, followed by 20 s perfusion with the same concentration of test alcohol in a barium frog Ringer's solution plus agonist solution. Flow rates for all channels were 2 mL/min, and were verified by fluid collection before each experiment. Return to pre-alcohol current values was evaluated 5 and 10 min after washout with barium frog Ringer's solution.

Clones for the human GABAAα1 and the rat GABAAβ2 subunits in pCIS II vectors were a gift from R. A. Harris (Institute for Cellular and Molecular Biology, Waggoner Center, University of Texas, Austin). Approximately 0.25–1 ng total plasmid, consisting of equal parts α1 and β2, were injected into the nucleus of frog oocytes that were studied 1–4 days later.

Amplifiers and data acquisition systems were identical to that described for NMDA studies. Alcohol solutions were prepared in glass syringes and administered to the bath at 2 mL/min using syringe pumps and polytetrafluoroethylene tubing. Transmembrane potentials were kept constant at –80 mV. Oocytes perfused with frog Ringer's solution for 5 min were switched to an identical solution containing 0.03 mM GABA for 20 s followed by 5 min of frog Ringer's solution; this was repeated 3 times to verify current reproducibility. A solution containing the test alcohol in frog Ringer's solution was then perfused for 1 min 40 s followed by the same alcohol concentration in frog Ringer's solution plus 0.03 mM GABA. Reversibility of alcohol effects was verified after 5 and 10 min of washout with frog Ringer's solution.

All solutions were prepared daily in either closed, sterile flexible containers (Intravia, Baxter, Round Lake, IL) or in gas tight glass syringes equipped with Teflon stoppers. Racemic alcohols were prepared from equal parts of each stereoisomer. The applied alcohol concentrations (16 mM 2-butanol; 4 mM 2-pentanol; and 2 mM 2-hexanol) are close to the blood alcohol concentrations measured at MAC in rats (2–11). Alcohol vapor concentrations in equilibrium with perfusate were determined at the end of each experiment using headspace gas chromatography with a Carbowax column (Series 580; Gow-Mac, Bethlehem, PA), from which solution concentrations were calculated using previously determined Ostwald saline-gas partition coefficients at room temperature (11).

The percent change in current for TRESK was calculated as the difference in the current change (−60 to + 60 clamp) after 2 min of alcohol exposure versus the minute immediately preceding alcohol exposure, divided by the pre-exposure current. The percent change for NMDA and GABA currents was calculated using the pre-post agonist current difference during alcohol exposure and immediately preceding alcohol exposure.

Responses to all isomers (R, S, and racemate) were evaluated each study day. When possible, alcohols were studied using a randomized cross-over design, with each oocyte serving as its own control. For these alcohols and receptors, data were analyzed using repeated-measures analysis of variance with Dunn-Sidak post hoc tests (SPSS, Chicago, IL). However, because of baseline drift during some prolonged studies, complete repeated measures were not always possible. For these data sets, analysis was performed using Student's t-test with Bonferroni corrections for multiple comparisons. Responses were significant at P < 0.05.


Alcohol responses for oocytes expressing TRESK, GABAA, and NMDA are shown in Table 1. Sample electrophysiologic recordings are shown in Figures 1–3. Current differences for TRESK before alcohol exposure and after alcohol washout were within 10%. Differences between baseline and washout peaks for both GABAA and NMDA currents were within 15%. Comparisons with water-injected oocytes verified that <2% of absolute responses were attributable to background currents.

Table 1:
Percent Change in Currents During Two-Electrode Voltage Clamp for Three Different Ion Channels to Three Chiral Alcohols
Figure 1.:
Two–electrode voltage clamp tracing from an oocyte expressing human TRESK channels, with current in microoamperes (μA) expressed as a function of time. 16 mM racemic 2-butanol, the concentration of 2-butanol at MAC, enhances currents through TRESK channels.
Figure 2.:
Two-electrode voltage clamp tracing from an oocyte expressing α1β2 GABAA receptors. Each peak corresponds to exposure to 30 μm GABA. The second peak shows enhanced currents in response to GABA when 1 MAC (16 mM) racemic 2-butanol was applied before and during GABA application. Currents are measured in nanoamperes (nA).
Figure 3.:
Sample tracing from an oocyte expressing NR1/NR2A NMDA receptors. Each peak corresponds to 100 μM glutamate plus 10 μM glycine exposure. Note the reduced current when the oocyte is exposed to 1 MAC (16 mM) racemic 2-butanol before and during agonist application.

Significant differences between 2-pentanol isomers were found for GABAA channels. Difference between 2-hexanol isomers were found for GABAA and NMDA receptors. The potencies of racemic alcohol mixtures were either intermediate between, or statistically indistinguishable from, one or both of their component stereoisomers on all channels studied and for all alcohols. No significant enantioselectivity was shown for TRESK with any alcohol.

Alcohol concentrations for the enantiomers, as determined by gas chromatography at the end of the experiment, differed by 10% or less, indicating that little alcohol was lost from the bag or apparatus during the course of these experiments.


We undertook this study to determine whether enantioselective interactions of anesthetic 2-alcohols on three structurally and functionally different anesthetic-sensitive ion channels paralleled those observed on MAC in rats. We found that they did not. Where enantioselectivity was observed for the three test channels, it roughly correlated with potency; this well known relationship for many drugs (i.e., potency predicts the strength of enantioselectivity) is called Pfeiffer's rule (12). However, chiral interactions on these three channels were largely discordant with those observed in animals, which did not follow Pfeiffer's rule. 2-Butanol was enantioselective for MAC, but not for any of the three ion channels; 2-pentanol was enantioselective for MAC but for only one of the three ion channels (GABAA receptors); 2-hexanol was not enantioselective for MAC but showed this pattern for two of the three ion channels (TRESK).

We can calculate the sensitivity and specificity of enantioselectivity as a test of relevance of molecular targets, assuming all three of the channels studied are relevant to MAC. Under this assumption, there was one true positive result, with GABAA receptors identified as being enantioselective for 2-pentanol. Enantioselectivity produced five false negative results: one for each channel with 2-butanol and one for TRESK channels and NMDA receptors for 2-pentanol. This gives a sensitivity of 1/6 or 16%. Again assuming all three channels are relevant, there was one true negative result (TRESK with 2-hexanol) and two false positive results (GABAA and NMDA receptors with 2-hexanol). This gives a specificity of 1/3 or 33%. This sensitivity and specificity are poor.

If all three channels studied are not relevant to anesthesia, enantioselectivity has a better sensitivity and specificity. Under this assumption, the previously classified true negatives and positives would now be false negatives and positives and, similarly, the false negatives and positives would be true negatives and positives. This gives enantioselectivity a sensitivity of 5/6 or 0.83, and a specificity of 2/3 or 0.67. Such specificity and sensitivity are still inadequate to use enantioselectivity as a test.

Given the current understanding of anesthetic mechanisms, the poor sensitivity and specificity of chirality as a test of relevance of anesthetic targets is not surprising. We can consider this from the perspective of the erroneous (i.e., false positive and negative) results we expect the test to produce. Let us examine the issue of false negative results first. It is generally thought that more than one protein target underlies most anesthetic effects including MAC (13), as is generally true of traits showing continuous variation (14). When enantioselectivity determines a behavior, it reflects the aggregate effect of the isomers on those molecular targets. Enantioselectivity need not be observed on every target to observe enantioselectivity in the animal, however. Thus false negative results are expected using enantioselectivity as a test: relevant targets that do not show enantioselectivity will give negative results when another target in the neural circuitry producing MAC is enantioselective and renders MAC enantioselective.

Next, consider the issue of false positive results. Anesthetics have multiple behavioral effects. We have suggested that anesthesia is a state of immobility (13) and amnesia (15); others argue additional effects are important measures of anesthesia, with loss of righting reflex being particularly frequently measured in rats and mice (16). However, all agree that anesthesia consists of more than one trait, served by differing neural circuits. Given this circumstance, it is entirely possible that positive results for stereoselectivity may be false positive results. One way this could happen would be if a molecular target enantioselectively mediated the amnestic effects of an anesthetic but was not in the circuitry mediating the immobilizing effects of the anesthetic. If the anesthetic were not enantioselective for MAC, in this example the molecular target would produce a false positive result.

Other erroneous results may also occur. For example, isoflurane in one study was enantioselective in its effects on MAC (4). It was also enantioselective in its effects on enhancement of GABAA receptors (17). However, since those studies were published, evidence has accumulated indicating that GABAA receptors probably do not mediate MAC (18–19). This suggests that channels other than GABAA receptors mediate the enantioselectiviy of isoflurane on MAC. The relationship between enantioselectivity observed in the animal and on GABAA receptors may thus be a spurious association.

Another possible reason for the lack of enantioselectivity of 2-hexanol in rats is suggested by the data in Table 1. On GABAA receptors, the R isomer of 2-hexanol has a greater inhibitory effect that the S isomer, whereas the opposite is true for the NMDA receptor. Possibly, these countervailing effects eliminate enantioselectivity in the animal.

The problems of false positive and false negative results are to be expected for anesthetics acting on multiple molecular targets, such as inhaled anesthetics. Anesthetics that act on only a single molecular target, such as etomidate (20), should not suffer from these problems. Similarly, genetic studies in which enantioselectivity on molecular targets is evaluated via studies in intact animals (21) will not have these problems.

Our study has limitations. We chose representatives from families of anesthetic-sensitive ion channels. Enantioselectivity may be idiosyncratic to particular members of those families or particular combinations of subunits, rather than shared by the channels we studied. In addition, our ion channels or their subunits were cloned from humans and mice as well as rats. The enantioselective effects of anesthetic alcohols on these channels were compared with those observed on MAC in rats. It may not be valid to use channels other than those of the rat substrain in which the animal studies were performed. In making these choices, we reasoned that the biophysical interaction producing enantioselectivity, for example the complementary (chiral) fit of a binding pocket for an anesthetic, would be similar between related ion channels that were sensitive to anesthetics.

We conclude that matching patterns of enantioselectivity in animals and on molecular targets to determine the relevance of those targets is problematic because of the inadequate sensitivity and specificity of this test. We recommend that enantioselectivity not be used as a test of relevance for inhaled anesthetic targets.


1. Dickinson R, White I, Lieb WR, Franks NP. Stereoselective loss of righting reflex in rats by isoflurane. Anesthesiology 2000;93:837–43.
2. Won A, Oh I, Laster MJ, et al. Chirality in anesthesia I: minimum alveolar concentration of secondary alcohol enantiomers. Anesth Analg 2006;103:81–4.
3. Eger EI II, Koblin DD, Laster MJ, et al. Minimum alveolar anesthetic concentration values for the enantiomers of isoflurane differ minimally. Anesth Analg 1997; 85:188–92.
4. Lysko GS, Robinson JL, Casto R, Ferrone RA. The stereospecific effects of isoflurane isomers in vivo. Eur J Pharmacol 1994;263:25–9.
5. Braunwald E, Isselbacher K, Petersdorf R. Harrison's principles of internal medicine. New York: McGraw-Hill Book Company, 1987.
6. Patel AJ, Honore E. Anesthetic-sensitive 2P domain K+ channels. Anesthesiology 2001;95:1013–21.
7. Liu C, Au JD, Zou HL, et al. Potent activation of the human tandem pore domain K channel TRESK with clinical concentrations of volatile anesthetics. Anesth Analg 2004;99:1715–22.
8. Mihic SJ, Ye Q, Wick MJ, et al. Sites of alcohol and volatile anaesthetic action on GABA(A) and glycine receptors. Nature 1997;389:385–9.
9. Hollmann MW, Liu HT, Hoenemann CW, et al. Modulation of NMDA receptor function by ketamine and magnesium. Part II: interactions with volatile anesthetics. Anesth Analg 2001;92:1182–91.
10. Raines DE, Gioia F, Claycomb RJ, Stevens RJ. The N-methyl-d-aspartate receptor inhibitory potencies of aromatic inhaled drugs of abuse: evidence for modulation by cation-pi interactions. J Pharmacol Exp Ther 2004;311:14–21.
11. Won A, Oh I, Liao M. MAC of 2-,3-, and 4-alcohols and ketones in rats: relevance to anesthetic mechanisms. Anesth Analg In press.
12. Pfeiffer CC. Optical isomerism and pharmacological action, a generalisation. Science 1956;124:29–31.
13. 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.
14. Falconer D, Mackay T. Introduction to quantitative genetics, 4th ed. London: Pearson Prentice Hall; 1996.
15. Sonner JM, Xing Y, Zhang Y, et al. Administration of epinephrine does not increase learning of fear to tone in rats anesthetized with isoflurane or desflurane. Anesth Analg 2005;100:1333–7.
16. Belelli D, Muntoni AL, Merrywest SD, et al. The in vitro and in vivo enantioselectivity of etomidate implicates the GABAA receptor in general anaesthesia. Neuropharmacology 2003;45:57–71.
17. Hall AC, Lieb WR, Franks NP. Stereoselective and non-stereoselective actions of isoflurane on the GABAA receptor. Br J Pharmacol 1994;112:906–10.
18. Sonner JM, Zhang Y, Stabernack C, et al. GABA(A) receptor blockade antagonizes the immobilizing action of propofol but not ketamine or isoflurane in a dose-related manner. Anesth Analg 2003;96:706–12.
19. Zhang Y, Sonner JM, Eger EI II, et al. Gamma-aminobutyric acidA receptors do not mediate the immobility produced by isoflurane. Anesth Analg 2004;99:85–90.
20. Jurd R, Arras M, Lambert S, et al. General anesthetic actions in vivo strongly attenuated by a point mutation in the GABA(A) receptor beta3 subunit. Faseb J 2003;17:250–2.
21. Sedensky MM, Cascorbi HF, Meinwald J, et al. Genetic differences affecting the potency of stereoisomers of halothane. Proc Natl Acad Sci U S A 1994;91:10054–8.
© 2006 International Anethesia Research Society