From 1900 until the 1970s–1980s, the Meyer-Overton hypothesis (1,2) dominated thinking concerning theories of narcosis. The Meyer-Overton hypothesis suggests that affinity for a lipid-like phase determines anesthetic potency and predicts that the product of some measure of lipophilicity (e.g., the lipid/gas partition coefficient) and some measure of anesthetic potency (e.g., MAC) equals a constant. MAC is the minimum alveolar concentration of anesthetic required to eliminate movement in response to a noxious stimulus in 50% of subjects. For conventional anesthetics, the olive oil/gas partition coefficient times MAC in rats equals 1.82 ± 0.56 atm (3).
The correlation of potency with lipophilicity suggested that anesthetics acted by an effect on a lipid-like site, perhaps a site within the membrane bilayer of neurons. A sea-change in thinking occurred in the 1970s and 1980s, largely because of the arguments by White et al. (4–7) and by Franks and Lieb (8,9) that an effect of anesthetics on a hydrophobic portion of proteins, perhaps receptors, could equally explain the correlation of potency and lipophilicity. This focus on proteins (e.g., ligand-gated or voltage-gated channels) continues to the present.
An additional factor led to an abandonment of purely lipid-based theories of narcosis. Certain lipophilic inhaled compounds were found to be far less potent than their lipophilicity would predict, which conflicted with the Meyer-Overton hypothesis. Indeed, some such compounds (the so-called nonimmobilizers) do not produce anesthesia (i.e., do not cause immobility) by themselves and do not add to the immobilization produced by known anesthetics (10–12). A low affinity to water (i.e., saline/gas partition coefficients usually <0.02) distinguishes these compounds from conventional anesthetics, and this “nonpolarity” has been suggested to explain their inability to produce anesthesia (13–16): Their nonpolar nature suggests that they cannot influence events at interfaces because they cannot achieve sufficient interfacial concentrations. At the other end of the spectrum, alcohols have potencies greater than might be predicted from the Meyer-Overton hypothesis by their lipophilicities (17,18). Alcohols have large saline/gas partition coefficients.
These relationships suggest that the inhaled anesthetic site of action has both aqueous (polar) and lipid-like (nonpolar) (i.e., amphipathic) characteristics (1821). Anesthesia requires that the anesthetic molecule have affinities for both phases. One might argue that the Meyer-Overton hypothesis, modified to accommodate the additional importance of hydrophilicity, has not been disproven (20). Cantor (22,23) proposes that anesthetics may act in phospholipid bilayers by distorting the lateral pressure in such bilayers. This might be particularly important at the membrane-water or membrane-protein interface. In support of this notion, long-chained polyhydric alcohols, compounds too long and large to fit completely into a protein “pocket”—and in carbon length far beyond what is thought to be a cutoff for alcohol potency—still can produce anesthesia (or partial anesthesia) in tadpoles (24). In the case of a sterically constrained binding site in a protein, these long molecules would fill the binding site and the remainder would “dangle” into the surrounding lipid or water phase. In contrast, long polyhydroxy molecules could be fully accommodated within the phospholipid-water interface of a membrane (24).
Among many others, three general mechanisms have been considered for the production of anesthesia by alcohols: binding in a water-filled cavity in proteins, binding at the water-lipid or protein-lipid interface of membranes, and partitioning into bulk lipid (25). In proteins, the hydroxyl group of the alcohol could anchor to the protein binding site by hydrogen bonding with a serine or threonine, and the carbon tail could interact with hydrophobic amino acids (26). An analogous mechanism could occur at the membrane interface where the polar headgroups of phospholipids could anchor to hydroxyl groups and the carbon chain could interact with the acyl chains. Both of these mechanisms predict that the number of carbons attached to the C-OH group rather than the total number of carbons would determine the potency of the alcohol. In contrast, solubility in bulk lipid would depend more on the total number of carbons and would not depend strongly on the position of the hydroxyl group. This lipid solubility can be measured experimentally as the olive oil/gas partition coefficient.
To address these possibilities, we measured the effect of C(???)OH placement in various alcohols and also the C=O group in ketones (which should act similarly to the C–OH) on the potency of those alcohols and ketones as defined by MAC. We proposed that the length of the carbon chain extending from the C(???)OH and C=O group, rather than the total carbon chain length, would define potency. For example, we hypothesized that 2-pentanol (CH3CH2CH2CHOHCH3) would have a potency closer to 1-butanol (CH3CH2CH2CH2OH) than to 1-pentanol (CH3CH2CH2CH2CH2OH).
With approval of the Committee on Animal Research of the University of CA, San Francisco, we studied 352 male specific-pathogen-free, Sprague-Dawley rats (Crl:CD(SD)BR) weighing 300–450 g obtained from Charles River Laboratories (Hollister, CA). These rats separately received various alcohols and ketones intraperitoneally, diluted in saline or olive oil. Data from additional rats given n-alcohols and reported previously (17,27) were added to the data gathered for the compounds tested in the present study. Each animal was caged with up to as many as two additional rats, and all had continuous access to standard rat chow and tap water before study. We studied eight alcohols and seven ketones (Table 1).
Determination of MAC of Alcohols and Ketones
Alcohols and ketones having four carbons [2-butanol and 2-butanone] were dissolved in saline, and a 10% solution was injected as two equal injections made 5 min apart. The remaining alcohols and ketones were dissolved in olive oil (2%–10%), and each of those solutions was injected intraperitoneally as two equal injections made 5 min apart.
One to four rats were prepared for each study. Immediately after injection of the test compound, we placed each rat in an individual plastic cylinder to which we delivered oxygen at a flow rate of 1 L/min/rat. We monitored temperature by a rectal probe and maintained temperature between 37°C and 39°C by application of heat.
At 40 min after the last alcohol or ketone injection, we stimulated the rat's tail by applying and moving an alligator clamp for up to 1 min and observing the rat for movement of an extremity or the head. Forty minutes was selected by trial and error in our previous studies of n-alcohols (17), and fortunately seems to be the time of peak concentration in the rat brain after intraperitoneal injection of ethanol (28). If no movement occurred, we made an incision in the abdomen to permit withdrawal of approximately 10 mL of blood from the aorta into a heparinized 10-mL syringe. If movement occurred, we administered isoflurane (1%–2%) in 1 L/min of oxygen to achieve anesthesia (lack of movement) and then obtained the blood sample. This sample was transferred to a 50-mL glass syringe.
To determine the partial pressure of the alcohol or ketone in aortic blood, we added 20–25 mL of air to the syringe and equilibrated the contents for 60 min in an incubator at 37°C. The gas phase was analyzed by gas chromatography for alcohol or ketone (see below), using primary volumetric standards. Because of the great solubility of alcohols and ketones in blood, this gas phase value immediately indicated the partial pressure in blood.
We applied a logistic regression analysis to the resulting data (29). Each rat supplied two values: the response (movement or no movement) and the associated alcohol or ketone partial pressure. The logistic regression analysis supplied a value for the partial pressure producing absence of movement in 50% of rats (the EC50) and the variance (standard error) about this value. We define this EC50 as MAC.
Correction of the MAC for Alcohols from Its Ketone Metabolite Contribution
In the studies of the MAC of the 2-, 3-, and 4-alcohols, a large contaminant peak appeared before the alcohol peak. The retention time of this peak indicated that it was the ketone metabolite of the alcohol. As might be expected from zero-order kinetics for the metabolism of alcohol, the ketone peak did not vary appreciably as a function of the concentration of the alcohol. That is, the ketone concentration was essentially constant. As indicated by our studies of the MAC of the ketones, the ketone metabolite contributed to the MAC produced by the parent alcohol. Assuming additivity, we corrected the alcohol MAC for the ketone metabolite contribution. We also determined that corresponding aldehydes did not result from the metabolism of the n-alcohols.
Measurement of the Solubilities of Alcohols and Ketones in Saline and Olive Oil
The test alcohols and ketones are highly soluble in olive oil, making it difficult to measure concentration in the solvent phase by extraction. Accordingly, for all but the least soluble alcohols, we added a small aliquot of alcohol or ketone to an aliquot of each of the solvents to derive a calculated solvent concentration. That is, knowing (or measuring) the density of each alcohol and ketone, we calculated the concentration of the alcohol in the solvent (CS). Approximately (because of the enormous solubility of these compounds in lipid-like material, the exact amount does not matter) 10 mL of the combination of alcohol or ketone and solvent was placed in 50-mL syringes, 30 mL of air was added, and the mixture was equilibrated at 37°C for at least 1 h in an incubator. After this time, the gas phase was analyzed for the concentration of the alcohol or ketone (CG). The solvent/gas partition coefficient was calculated as CS/CG.
The same method was applied to the determination of the saline/gas partition coefficients for the lower molecular weight alcohols and ketones. For larger alcohols and ketones, the saline/gas partition coefficient was determined using a modification of techniques described previously (30).
Gas Chromatographic Analysis
We used a Gow-Mac 580 or 750 flame ionization detector gas chromatograph (Gow-Mac Instrument Corp., Bridgewater, NJ). The 4.6- or 9.2-m, 0.22-cm (id) column was packed with carbowax. Column temperature was adjusted to give reasonable (5–15 min) retention times. The detector was maintained at temperatures 30–50°C warmer than the column. The carrier flow was nitrogen. The detector received 36–45 mL/min hydrogen and 200–300 mL/min air.
Tables 1–4 provide data from our current studies and from our previously published studies for the n-alcohols from ethanol to octanol (17). Table 1 provides some physical properties and the sources of the tested alcohols and ketones. Table 2 provides the saline/gas and oil/gas solubilities for the test compounds. Table 3 provides MAC values and the product values (of the oil/gas partition coefficient times the MAC) for the tested ketone compounds. Table 4 provides two MAC values (one for before the correction and one for after the correction) and the product values (of the oil/gas partition coefficient times the corrected MAC) for the tested alcohol compounds. For the product of MAC times the olive oil/gas partition coefficient, the alcohols and ketones disobeyed the Meyer-Overton hypothesis. The ketones were slightly less than twice as potent as might be predicted from their oil/gas partition coefficients (Table 3), whereas the alcohols were approximately 10 times more potent than expected using comparison values obtained previously for conventional anesthetics (Table 4). MAC for both ketones and alcohols correlated inversely with the largest number of carbons from the oxygen (Fig. 1), the number of carbons (Fig. 2), and the oil/gas partition coefficient (Fig. 3). In all such correlations, the values for ketones lay above the values for the alcohols.
As we proposed, the point of attachment of the oxygen in the alcohols considerably influenced potency (Table 4). For example, MAC for 2-pentanol is close to that of 1-butanol; MAC for 2-hexanol is close to that of 1-pentanol. However, this relationship does not apply to the ketones (Table 3). For example, MAC for 3-hexanone is half that of 2-pentanone; MAC for 4-octanone is one-third that of 2-hexanone.
We had anticipated that potency might best correlate with the length of the longest carbon chain that extends from the point of oxygen attachment for either the alcohols or the ketones. We envisioned that the oxygen tethered the molecule at an interface between the hydrophobic and aqueous portions of the site of anesthetic action and that the extent of the intrusion of the carbon chain into the hydrophobic portion might determine potency (31). Indeed, there is a reasonable correlation for the alcohols and separately for the ketones (Fig. 1). However, the correlation is no better than the correlation simply with the total length of the carbon chain, regardless of placement of the oxygen (Fig. 2). The only advantage of the correlation seen in Figure 1 is that it brings the ketone and alcohol values closer together. Finally, neither of the preceding correlations improves on the correlation between potency (MAC) and lipophilicity (the oil/gas partition coefficient) (Fig. 3). The increase in potency for both the alcohols and ketones with increasing chain length and lipophilicity has been seen in many studies of series of halogenated and unhalogenated hydrocarbons (17,18,32,33).
One limitation to our argument—that the longer carbon chain that extends from the C-OH group determines potency—is the failure to account for the effect of the residual shorter chain. The shorter chain might contribute to potency in an uncertain manner, and might increasingly become a factor as it lengthens (i.e., it might become more important in 3-heptanol than in 2-heptanol).
Although the correlation proposed by Meyer (1) and Overton (2) applies separately to the ketones and alcohols (Fig. 3), the ketone correlation lies above that for the alcohols. As suggested previously, hydrophilicity as well as lipophilicity influence potency (13,14,18,19,21,34–36). The saline/gas partition coefficients for the ketones are approximately a fifth as great as for alcohols with the same chain length and attachment of oxygen (Table 2). Perhaps this explains the position of the data for ketones versus alcohols (Figs. 1–3). However, Janoff et al. (37) might remind us that we used a solvent (olive oil) that may inadequately represent the site of anesthetic action; they might argue that phosphatidylcholine bilayers present a better model for alcohols, conventional inhaled anesthetics, and various injected anesthetics, including barbiturates.
Whereas they have essentially identical molecular weights and volumes, the key difference between the alcohols and ketones in Figure 3 is that the alcohols can both donate and accept hydrogen bonds, whereas ketones can only accept them. The greater potency of the alcohols suggests the importance of hydrogen bond donation by the alcohols. This conclusion is supported by previous studies predicting that an anesthetic binding site should accept a hydrogen bond about as well as water (21).
Additional support for the importance of hydrogen bonds is provided by the boiling point data in Table 1. The boiling point of the primary alcohols is 10ºC–20ºC higher than the corresponding secondary or tertiary alcohols. Elevation of boiling points provides strong evidence for inter-molecular hydrogen bonds (38). For the alcohols in Table 1, the primary alcohols are clearly able to form stronger intermolecular bonds. We tested the possibility that hydrogen bonding is important for MAC by graphing separately primary and secondary/tertiary alcohols versus MAC. The primary alcohols are statistically more potent than the secondary/tertiary alcohols and the regression lines separate into two parallel graphs.
Several of the test alcohols used in the present study (those having the C-OH group on a carbon other than a terminal carbon) have chiral centers. We tested the anesthetic effect of racemic mixtures of such alcohols. Thus, it might be possible that differences between the isomers might confound the implications of our findings to our hypotheses. However, we also have tested the differences in MAC of the isomers (unpublished data) and find but limited differences (at most a doubling) in MAC. We argue that because such differences are limited, they should not compromise our interpretations of the data.
We conclude that our results seem to be most consistent with a bulk lipid-partitioning model and show that the ability to donate a hydrogen bond increases potency. Our data do not support a simple model in which the polar group is anchored at a protein-lipid or water-lipid interface and controls the extension of the alkane chain into a site of defined size. However, the differences in correlation coefficients are too small to exclude either a protein site model or a membrane surface model.
1. Meyer HH. Theorie der Alkoholnarkose. Arch Exptl Pathol Pharmakol 1899; 42:109–18.
2. Overton E. Studien über die Narkose, Zugleich ein Beitrag zur allgemeinen Pharmakologie. Jena, Germany: Gustav Fischer, 1901:1–195.
3. Taheri S, Halsey MJ, Liu J, et al. What solvent best represents the site of action of inhaled anesthetics in humans, rats and dogs? Anesth Analg 1991;72:627–34.
4. White DC, Dundas CR. The effect of anaesthetic agents on the emission of light by luminous bacteria. Br J Anaesth 1969;41:194.
5. White DC, Dundas CR, Adey G. The site of action of anaesthetic agents: post-anoxic flashes in luminous bacteria. Br J Anaesth 1971;43:716.
6. White DC, Dundas CR. Effect of anaesthetics on emission of light by luminous bacteria. Nature 1970;226:456–8.
7. White DC, Dundas CR. Potency of anaesthetic agents measured by luminous bacteria. Br J Anaesth 1970;42:89.
8. Franks NP, Lieb WR. Do general anaesthetics act by competitive binding to specific receptors? Nature 1984;310:599–601.
9. Franks N, Lieb W. Molecular mechanisms of general anaesthesia. Nature 1982;300:487–93.
10. Koblin DD, Chortkoff BS, Laster MJ, et al. Polyhalogenated and perfluorinated compounds that disobey the Meyer-Overton hypothesis. Anesth Analg 1994;79:1043–8.
11. Liu J, Laster MJ, Koblin DD, et al. A cut-off in potency exists in the perfluoroalkanes. Anesth Analg 1994;79:238–44.
12. Eger EI 2nd, Liu J, Koblin DD, et al. Molecular properties of the ‘ideal’ inhaled anesthetic: Studies of fluorinated methanes, ethanes, propanes and butanes. Anesth Analg 1994;79:245–51.
13. Pohorille A, Cieplak P, Wilson MA. Interactions of anesthetics with the membrane-water interface. Chem Phys 1996;204:337–45.
14. Eger EI 2nd, Koblin DD, Harris A, et al. Hypothesis: Inhaled anesthetics produce immobility and amnesia by different mechanisms at different sites. Anesth Analg 1997;84:915–8.
15. Chipot C, Wilson MA, Pohorille A. Interactions of anesthetics with the hexane-water interface. J Phys Chem B 1996;101:782–91.
16. Pohorille A, Wilson MA, Cieplak P. Interaction of alcohols and anesthetics with the water-hexane interface: A molecular dynamics study. Prog Colloid Polym Sci 1997;103:29–40.
17. Fang ZX, Ionescu P, Chortkoff BS, et al. Anesthetic potencies of n-alkanols: Results of additivity studies suggest a mechanism of action similar to that for conventional inhaled anesthetics. Anesth Analg 1997;84:1042–8.
18. Eger EI 2nd, Ionescu P, Laster MJ, et al. MAC of fluorinated alkanols in rats: Relevance to theories of narcosis. Anesth Analg 1999;88:867–76.
19. Hansch C, Vittoria A, Silipo C, Jow PY. Partition coefficients and the structure-activity relationship of the anesthetic gases. J Med Chem 1975;18:546–8.
20. Sonner JM, Antognini JF, Dutton RC, et al. Inhaled anesthetics and immobility: Mechanisms, mysteries, and MAC. Anesth Analg 2003;97:718–40.
21. Abraham MH, Lieb WR, Franks NP. Role of hydrogen bonding in general anesthesia. J Pharm Sci 1991;80:719–24.
22. Cantor RS. Lateral pressures in cell membranes: A mechanism for modulation of protein function. J Phys Chem B 1997;101:1723–5.
23. Cantor RS. The lateral pressure profile in membranes: A physical mechanism of general anesthesia. Biochemistry 1997;36:2339–44.
24. Mohr JT, Gribble GW, Lin SS, et al. Anesthetic potency of two novel synthetic polyhydric alkanols longer than the n-alkanol cutoff: evidence for a bilayer-mediated mechanism of anesthesia? J Med Chem 2005;48:4172–6.
25. Trudell JR, Harris RA. Are sobriety and consciousness determined by water in protein cavities? Alcohol Clin Exp Res 2004;28:1–3.
26. Kruse SW, Zhao R, Smith DP, Jones DN. Structure of a specific alcohol-binding site defined by the odorant binding protein LUSH from Drosophila melanogaster. Nat Struct Biol 2003;10:694–700.
27. Fang ZX, Gong D, Ionescu P, et al. Maturation decreases ethanol MAC more than desflurane MAC in rats. Anesth Analg 1997;84:852–8.
28. Yoshimoto K, Komura S. Monitoring of ethanol levels in the rat nucleus accumbens by brain microdialysis. Alcohol Alcoholism 1993;28:171–4.
29. Waud DR. On biological assays involving quantal responses. J Pharmacol Exper Therap 1972;183:577–607.
30. Taheri S, Laster MJ, Liu J, et al. Anesthesia by n-alkanes not consistent with the Meyer-Overton hypothesis: Determinations of the solubilities of alkanes in saline and various lipids. Anesth Analg 1993;77:7–11.
31. Moss GWJ, Curry S, Franks NP, Lieb WR. Mapping the polarity profiles of general anesthetic target sites using n-alkane-(α,ω)-diols. Biochemistry 1991;30:10551–7.
32. Eger EI 2nd, Laster MJ. The effect of rigidity, shape, unsaturation, and length on the anesthetic potency of hydrocarbons. Anesth Analg 2001;92:1477–82.
33. Liu J, Laster MJ, Taheri S, et al. Effect of n-alkane kinetics in rats on potency estimations and the Meyer-Overton hypothesis. Anesth Analg 1994;79:1049–55.
34. Koblin DD, Laster MJ, Ionescu P, et al. Polyhalogenated methyl ethyl ethers: Solubilities and anesthetic properties. Anesth Analg 1999;88:1161–7.
35. Pohorille A, Wilson MA. Excess chemical potential of small solutes across water-membrane and water-hexane interfaces. J Chem Phys 1996;104:3760–73.
36. Katz Y, Simon S. Physical parameters of the anesthetic site. Biochem Biophys Acta 1977;471:1–15.
37. Janoff AS, Pringle MJ, Miller KW. Correlation of general anesthetic potency with solubility in membranes. Biochem Biophys Acta 1981;649:125–8.
38. Pauling L. The nature of the chemical bond, 3rd ed. Ithica, NY: Cornell University Press, 1960: 499.