Volatile Anesthetic Additivity and Specificity in Saccharomyces cerevisiae : Implications for Yeast as a Model System to Study Mechanisms of Anesthetic Action

All strains were derived from the haploid Zzz⁺(wild-type) yeast RLK88–3C (MATa his4–260 leu2–3,112 ura3–52 ade2–1 trp1-HIII lys2-Delta BX can1^R). [17]Strain CJP117–2BF (zzz1 Delta) has a genomic deletion of ZZZ1 (Genbank Accession numbers Z49269 and Z49704) from bases -25 to +2895 (HincII and EcoRV restriction sites) with respect to the translation initiation codon (Keil and Wolfe, unpublished data). CJP113–2B (zzz4 Delta) has a genomic deletion of ZZZ4 (Genbank Accession Z28213), as described. [9]Synthetic complete media lacking various amino acids or nucleic acid bases were prepared as described previously. [17]Synthetic minimal medium is similar to synthetic complete medium except it contains only the amino acids and nucleic acid bases required by a strain. Plasmids were introduced into yeast strains by lithium acetate transformation [18]and selected on synthetic medium lacking uracil (synthetic complete-ura). Cycloheximide (Sigma Chemical Co., St. Louis, MO) was dissolved in dimethyl sulfoxide at 14.2 mM and added to synthetic complete-ura medium to a final concentration of 2.1 [micro sign]M just before the plates were poured. Cadmium sulfate (Sigma Chemical Co.) was dissolved in water at 0.5 M, filter sterilized and added to synthetic complete-ura medium to final concentrations of 15 and 30 [micro sign]M. Sulfometuron methyl (provided by Du Pont Agricultural Products Department, Wilmington, DE) was dissolved in dimethyl sulfoxide at 11 mM and added to synthetic minimal medium to final concentrations of 5.5 and 16.4 [micro sign]M. Chloramphenicol (Sigma Chemical Co.) was dissolved in ethanol at 0.62 M and added to YEPG [19]media to final concentrations of 3.1 and 9.3 mM. For drugs dissolved in dimethyl sulfoxide or ethanol, an equal concentration of the carrier was added to media containing or lacking drug.

Plasmid DNA was propagated in Escherichia coli and isolated using the Wizard Minipreps DNA purification system (Promega, Madison WI). Plasmid pDR3.3, [20]which overexpresses PDR5, was obtained from John Golin. Plasmid pSEY18-R2.5, [21]which overexpresses YAP1/PDR4, was obtained from W. Scott Moye-Rowley. Yep24, [22]the control plasmid, was obtained from D. Botstein.

To assess the reaction of strains to various drugs, freshly saturated cultures, grown in appropriate liquid media to maintain any resident plasmids, were diluted 1:50 in sterile glass distilled water and 5 [micro sign]l was spotted on appropriate solid media.

For additivity studies, the concentration of one agent was held constant at 0.25, 0.50, or 0.75 MIC and the minimum concentration of the second agent, varied at 0.05-MIC increments, that prevents visible growth for 3 days, was determined. Environments of volatile anesthetics were achieved by injecting liquid agents into partially evacuated, sealed chambers. Concentrations of individual agents (+/- 0.02 MIC) were determined as described previously using three to five independent samples from a chamber. [9]The concentration of volatile agents decreased by <0.05 MIC during the 3-day exposure period. To produce atmospheres containing combinations of volatile agents, volumes of the liquid agents necessary to produce the desired partial MIC concentrations were sequentially injected into the partially evacuated, sealed chambers.

The effectiveness of volatile nonanesthetics as yeast growth inhibitors was assayed in a similar manner. From previous studies, the average value of the MIC:MAC ratio is 8.76. [9]Based on the oil-gas partition coefficients of 1,2-dichlorohexafluorocyclobutane and 2,3-dichlorooctafluorobutane, [12]the human MAC can be predicted based on the assumption that MAC X (oil-gas partition coefficient)= 1.30. [23]The predicted human MACs are 0.029 and 0.052 atm, respectively. Thus the predicted MICs of these compounds are 0.25 and 0.46, respectively. Because these concentrations are above the saturated vapor pressures of the respective compounds, additivity studies using these nonanesthetic agents combined with isoflurane or methoxyflurane were performed. Strains were exposed as previously described [9]to a combination of saturated nonanesthetic plus 0.75 MIC conventional anesthetic, yielding additive predicted MIC values of 1.71 and 1.40 for 1,2-dichlorohexafluorocyclobutane and 2,3-dichlorooctafluorobutane, respectively.

Plates containing cycloheximide, cadmium, chloramphenicol, or sulfometuron methyl were incubated at 30 [degree sign]C for 2–3 days and scored for growth.

Isoflurane was provided by Anaquest (Liberty Corner, NJ). Halothane without thymol as a preservative was provided by Halocarbon Laboratories (North Augusta, SC). Methoxyflurane was provided by Abbott Laboratories (King of Prussia, PA). 1,2-Dichlorohexafluorocyclobutane (c[CCIFCCIFCF²CF²] >97%, lot 94–2650) and 2,3-dichlorooctafluorobutane (CF³CC1FCC1FCF³>97%, lot 95–0683) were purchased from PCR Inc. (Gainesville, FL).

To test volatile anesthetic additivity, pairwise combinations of agents were assayed for their ability to inhibit the growth of Zzz⁺yeast. In these experiments, methoxyflurane was set at a concentration of 0.25, 0.50, or 0.75 MIC (1.0 MIC for methoxyflurane = 0.0137 atm) and the isoflurane concentration was varied to yield the final MIC values shown in Table 1. As illustrated in Figure 1, which gives examples, and in Table 1and Figure 2, which show all the results with methoxyflurane and isoflurane, an additive concentration of at least 1.0 MIC was always needed to inhibit yeast growth completely. Similar studies were done using combinations of methoxyflurane plus halothane and isoflurane plus halothane. In all cases, additive concentrations of 1.0 MIC completely inhibit growth of our wild-type yeast strain, but some growth occurs at 0.95 MIC and lower (data not shown). Although pairs of methoxyflurane, halothane, and isoflurane display additivity for growth inhibition of Zzz⁺yeast, growth of the anesthetic-resistant mutants zzz1 Delta and zzz4 Delta is not inhibited by these anesthetics either singly or in combination when the summed concentrations are 1.0 MIC (Figure 1).

Two volatile, lipophilic, polyhalogenated compounds, 1,2-dichlorohexafluorocyclobutane and 2,3-dichlorooctafluorobutane, which do not produce anesthesia or display additivity with known anesthetics in rats, [12]were tested for growth inhibitory effects in yeast. Zzz⁺strains grew in an atmosphere saturated with either of these volatile nonanesthetic compounds (data not shown). This growth might be expected because the saturated vapor pressures of 1,2-dichlorohexafluorocyclobutane and 2,3-dichlorooctafluorobutane yield 0.96 and 0.65 of their respective predicted MICs. To further evaluate these nonanesthetics for growth inhibitory properties, additivity studies combining a nonagent and a known anesthetic were performed. Compared with growth in an atmosphere of 0.75 MIC (0.0103 atm) methoxyflurane, growth of Zzz⁺strains and Zzz^-mutants is not noticeably affected when they are grown in an atmosphere containing a combination of saturated nonanesthetic and 0.75 MIC methoxyflurane, yielding an additive predicted MIC of 1.71 for 1,2-dichlorohexafluorocyclobutane and 1.40 for 2,3-dichlorooctafluorobutane (Figure 3). Similar results were found combining isoflurane as the anesthetic with saturated concentrations of either non-anesthetic (data not shown). Thus, in agreement with what is found in rats, these two volatile polyhalogenated compounds lack the effects of conventional anesthetics in the yeast growth assay.

Control experiments showed that growth of the Zzz⁺strain is inhibited by 1.0 MIC methoxyflurane or 1.0 MIC (0.120 atm) isoflurane in the presence of saturated concentrations of 1,2-dichlorohexafluorocyclobutane or 2,3-dichlorooctafluorobutane (data not shown), indicating that these nonanesthetics do not noticeably antagonize the activity of anesthetics. As expected, Zzz^-mutants did grow in the environment containing these mixtures of anesthetics and nonanesthetics.

Because our mutant Zzz^-strains are resistant to all the volatile anesthetics tested, it was possible that these mutations rendered yeast resistant to many compounds. Such cross-resistance to drugs is reminiscent of pleiotropic drug resistance (Pdr^-) mutants of yeast that are resistant to many (nearly 30 are known) structurally and pharmacologically unrelated drugs. [14]Most yeast Pdr^-strains are resistant to cycloheximide, sulfometuron methyl, chloramphenicol, or all three, whereas some (as when YAP1/PDR4 is overexpressed) are also resistant to the heavy metal cadmium. The zzz1 Delta and zzz4 Delta strains showed the same level of sensitivity to cycloheximide, sulfometuron methyl, and chloramphenicol as Zzz⁺strains and were more sensitive to cadmium than control Zzz⁺strains (Figure 4). Thus anesthetic resistance does not appear to confer a Pdr^-phenotype.

In the converse experiment, yeast Pdr^-mutants were tested for their resistance to volatile anesthetics. Multicopy (YEp) plasmids containing either PDR5 or YAP1/PDR4 were separately introduced into our Zzz (⁺) strain to overexpress these genes. As expected, strains containing either the PDR5 or the YAP1/PDR4 overexpression plasmid are resistant to cycloheximide and sulfometuron methyl. Strains containing the YAP1/PDR4 plasmid are also resistant to cadmium (Figure 4), whereas strains containing the PDR5 plasmid are resistant to chloramphenicol (data not shown). However, none of these strains are more resistant to isoflurane than the Zzz⁺yeast (Figure 4). Thus the anesthetic-resistant Zzz^-mutants appear to be specifically resistant to volatile anesthetics and appear unrelated to the known pleiotropic drug resistance mutants of yeast.

In these yeast studies with volatile anesthetics, growth inhibition was chosen as a general end point to examine because disruption of any of various essential cellular activities may make the cells unable to grow. Some of the most obvious include signal transduction, DNA synthesis, transcription, translation, and nutrient transport. Growth inhibition is also readily assayed either qualitatively or quantitatively and mutants resistant to the growth inhibitory effects can be readily isolated, [9]factors that are important for the molecular genetic studies of volatile anesthetic action we are pursuing.

Additive effects of volatile anesthetics have long been recognized in mammalian systems, and unitary theories of narcosis have been proposed to explain this phenomenon. [23]Such theories postulate a single molecular structure, perhaps present in several different proteins or locations, that interacts with anesthetic molecules to produce their effects. These theories propose it is the number, not the chemical identity, of anesthetic molecules present at the proposed binding site(s) that leads to anesthesia by volatile agents. The additive effect of anesthetics on yeast growth (described previously) as well as other unpublished data from our studies are consistent with these theories. However, a common mechanism (pathway) does not necessarily equate to a common site of binding because multiple sites of anesthetic binding, each containing the anesthetic-sensitive molecular structure, may converge on a single pathway to produce anesthesia. Such a converging pathway, initiated at distinct anesthetic binding sites, has been proposed based on genetic studies with the nematode Caenorhabditis elegans. [24]Elucidating the initial steps in the ZZZ pathway may identify the nature of the anesthetic-sensitive molecular structure, whereas characterizing the final steps of the ZZZ pathway may delineate the mechanism directly responsible for growth inhibition.

The specificity of the response of yeast to volatile anesthetics has been examined in two ways. First, Zzz⁺strains, which are sensitive to volatile anesthetics in accordance with the Meyer-Overton relation, were tested for sensitivity to volatile lipophilic compounds that fail to anesthetize rats. These experiments tested whether growth inhibition is a specific response to anesthetics or simply a response to volatile lipophilic compounds regardless of their activity as anesthetics. Second, Zzz^-mutants, which are resistant to volatile anesthetics, were tested for resistance to compounds involved in pleiotropic drug resistance of yeast.

Koblin et al. [12]characterized volatile, lipophilic polyhalogenated compounds devoid of anesthetic activity in rats even at partial pressures two times the predicted MAC. These nonanesthetics also are not effective inhibitors of yeast growth even in additive concentrations well in excess of the predicted yeast MIC. Thus neither yeast growth inhibition nor anesthesia is due solely to exposure to volatile, lipophilic chemicals. That growth inhibition of yeast is produced by volatile, lipophilic, polyhalogenated compounds that are anesthetic in vertebrates but not by similar compounds that are nonanesthetic in vertebrates suggests that volatile anesthetics may be working through similar mechanisms in both systems.

Several chemical classes of volatile polyhalogenated hydrocarbons can inhibit yeast growth. [9]These include alkane and ether derivatives (halothane and isoflurane), linear and branched compounds (methoxyflurane and sevoflurane), and compounds halogenated with chlorine, fluorine, bromine, or all three. The nonanesthetics used in this study are similar to the anesthetic compounds with respect to (1) the types of halogen atoms bonded to carbons;(2) oil-gas partition coefficients; and (3) molecular size. However, a critical difference may be that the nonanesthetic compounds are completely halogenated and thus lack a carbon hydrogen bond, whereas all anesthetics tested thus far contain carbon bonded to hydrogen. Koblin et al. [12]found a decrease in anesthetic potency as hydrogenation is decreased in a series of n-alkanes. Thus, for this class of anesthetics there appears to be a requirement for partial hydrogenation for both yeast growth inhibition and mammalian anesthesia.

Pleiotropic drug resistance in yeast arises from over-expression or increased activity of one or more transmembrane multidrug transport proteins. [14]These multidrug pumps belong to a superfamily of ATP-binding transporters that includes proteins involved in mammalian multidrug resistance. [15]One model views these transporters as “hydrophobic vacuum cleaners” that expel lipophilic drugs as they enter the plasma membrane, [15,25]suggesting the possibility that Zzz^-and Pdr (^-) mutants may affect the same pathways. The zzz1 Delta and zzz4 Delta anesthetic-resistant mutants are sensitive to the PDR drugs cycloheximide, sulfometuron methyl, chloramphenicol, and the metal cadmium. All other Zzz (^-) mutants tested to date are also sensitive to these compounds (Wolfe and Keil, unpublished data). As expected, [14]overexpression of PDR5 or YAP1/PDR4 in Zzz⁺yeast yields strains resistant to cycloheximide or sulfometuron methyl. Neither PDR5 nor YAP1/PDR4 overexpression produces anesthetic resistance, indicating that Zzz^-mutants are distinct from these pleiotropic drug resistance genes of yeast. Furthermore, none of the four ZZZ genes cloned thus far have been previously isolated in extensive Pdr (^-) mutant hunts. Therefore, the ZZZ genes appear unrelated to the pleiotrophic drug resistance phenotype.

There are a remarkable number of parallels between the activity of these volatile compounds as yeast growth inhibitors and mammalian anesthetics. Our analysis shows that (1) yeast growth inhibition by volatile anesthetics obeys the Meyer-Overton relation [9];(2) growth inhibition is rapid and possibly occurs instantaneously but certainly in <15 min[section](3) growth inhibition is reversible [9];(4) the dose-response curve for inhibition of yeast growth is very sharp (illustrated in Figure 1), as it is for anesthetic-dose response in mammals [26];(5) concentrations of mixtures of agents required to inhibit yeast growth are additive;(6) yeast growth is not inhibited by volatile, lipophilic nonanesthetic compounds; and (7) Zzz^-mutants, which are cross-resistant to all tested volatile anesthetics, are distinct from yeast pleiotropic drug resistance genes.

One difference between the yeast and mammalian systems is that the concentration of volatile agent required to completely inhibit growth is approximately nine times higher than the concentration necessary to induce anesthesia. But the significance of this difference is not clear. In general, when compared with mammalian systems, yeast is fairly resistant to drugs. [27]Yeast may have permeability barriers to drugs presented by the cell wall or the cell membrane [27]or altered transport of drugs into or out of the cell. One or more of these factors could limit the concentration of volatile anesthetic present at its site(s) of action in yeast. In addition, the end point measured in yeast, inhibition of growth of >99.99% of the cells, is different than the end point measured in humans, immobility of 50% of patients.

Whatever the significance of the quantitative differences in anesthetic requirements for the assays used in yeast and humans, the many qualitative parallels suggest the exciting possibility that the targets or mechanisms (or both) of action of volatile anesthetics in these two systems are related.

The authors thank Drs. Ross Shiman, Anita Hopper, David Larach, Bosseau Murray, and Reeta Prusty for their interest and critical reading of the manuscript, and Mark Pizzini and Ryan Romeo for assistance with initial PDR experiments.

[section] Only 10% of unbudded cells bud in the presence of anesthetic, [13]and control experiments show that 25% of unbudded cells initiate buds within 15 min when incubated at 30 [degree sign]C in the absence of anesthetic.