Halothane and Isoflurane Alter the Calcium sup 2+ Binding Properties of Calmodulin

Bovine brain calmodulin was obtained from Sigma Chemical (St. Louis, MO) and Calbiochem (La Jolla, CA) and purity was confirmed by gel electrophoresis. The fluorescence emission of calmodulin at 320 nm was measured over a range of Calcium²+ concentrations (10 sup -7 - 10 sup -4 M) in the presence and absence of halothane and isoflurane. Solutions of fixed Calcium²+ concentrations were formulated according to calculations based on the program of Fabiato and Fabiato [14]as modified by Berman.* Halothane and isoflurane concentrations from 0.5-8.7% (0.25-5 mM) were generated by adding aliquots of anesthetic-saturated buffer to the experimental solutions. Anesthetic concentrations were verified by gas chromatography.

The intrinsic tyrosine fluorescence of calmodulin was measured in a spectrofluorometer (SLM-Aminco 8000, Rochester, NY) at an excitation wavelength of 280 nm and an emission wavelength of 320 nm. The fluorescence of calmodulin is known to increase on incubation with Calcium sup 2+. In control titration experiments, small aliquots of CaCl²were added to a fluorescence cuvette containing 1 micro Meter calmodulin, 50 mM hydroxyethylpiperazineethane sulfonic acid, 100 mM KCl, and 2 mM ethyleneglycol-bis-(beta-aminoethyl ether) tetraacetic acid, at 37 degrees Celsius and pH 7.0. Over the range of the titration, pH varied less than 0.25 units. This variation has been shown not to significantly alter the affinity of calmodulin for Calcium²+. [8]Titrations in the presence of anesthetic were carried out with known concentrations of halothane or isoflurane added from saturated solutions to cuvettes containing 2 ml buffer; the cuvettes were sealed with polytetrafluorethylene to maintain constant volatile anesthetic concentrations. Aliquots of CaCl²were added with a Hamilton syringe through the seal. In the reversibility experiments, the control and anesthetic additions were performed as described above, except that samples were kept in the dark after 15 min of exposure, and the cuvettes were unsealed and stirred for 1 h in a Nitrogen²atmosphere before CaCl²aliquots were added. This procedure allowed time for the anesthetics to evaporate from the calmodulin-containing solutions. The absence of anesthetic was verified by gas chromatography. The cuvettes were then resealed, and Calcium²+ titrations were carried out. Final concentrations of anesthetic were confirmed by gas chromatography.

The fluorescence data were normalized as follows. The fluorescence of calmodulin before CaCl²additions (F⁰) was subtracted from intrinsic tyrosine fluorescence of calmodulin (F), and the difference F - F⁰was divided by the maximum fluorescence (F^max) achieved after saturation with Calcium²+. The result, Fⁿ, equal to (F - F⁰)/F^max, was plotted versus Calcium²+ concentration. All titration data were fit by nonlinear regression to the Hill equation, Fⁿ= [Ca²+]ⁿ/([Ca²+]ⁿ+ K^dsup n). Parameters derived from nonlinear regression analyses were used with the Enzfitter program (R. J. Leatherbarrow, Elsevier Biosoft, Cambridge, UK) to generate the curves shown in Figure 1, Figure 2, Figure 3and Figure 4. The statistical significance for the differences between anesthetic treatment and control were determined by two-way analysis of variance with multiple comparisons by using the Student-Neuman-Keuls method. [15]P values less than 0.05 were considered statistically significant.

Calcium²+ binding was measured by the change in intrinsic tyrosine fluorescence of calmodulin. The change in fluorescence was plotted as a function of Calcium²+ and fit by nonlinear regression analysis to the Hill equation to yield a value for K^dand a value for n. Calmodulin from two sources was used in the experiments: calmodulin purchased from Sigma Chemical (CaM^s) and calmodulin purchased from Calbiochem (CaM^c). These two preparations were characterized by different K^dand n values (Table 1), a difference that we attribute to greater Calcium²+ binding to CaM^sthan to CaM^cbefore the Calcium²+ titration and fluorescence measurement. Despite this quantitative difference, the qualitative changes in Calcium²+ binding observed on exposure to halothane and isoflurane were essentially identical. The fluorescence spectra of Calcium²+ -bound calmodulin at low and high anesthetic concentrations were identical to the spectra of control samples.

Volatile anesthetics altered the Calcium²+ -binding affinity of CaM^sin a dose-dependent, biphasic fashion. Titrations were carried out at halothane concentrations of 0.57% (0.25 mM), 1.14% (0.5 mM), 2.27% (1.0 mM), and 5.68% (2.5 mM) and at isoflurane concentrations of 1.74% (0.66 mM) and 8.68% (3.3 mM). At low concentrations of halothane and isoflurane, the affinity of calmodulin for Calcium²+ was decreased. At 1.74% isoflurane, for example, K^dwas increased to 6.99 x 10 sup -7 from 5.89 x 10 sup -7 M, indicating lower affinity. However, at higher concentrations of both anesthetics, the affinity for Calcium²+ was increased. In the presence of 8.68% isoflurane, the K^dfor Calcium²+ decreased from 5.89 x 10 sup -7 to 5.09 x 10 sup -7 M (Figure 1and Figure 2and Table 1).

When Calcium²+ titrations of CaM^swere carried out in the presence of 0.57% halothane and 1.74% isoflurane, the Calcium²+ -calmodulin binding curve shifted to the right. These changes were statistically significantly different from control at the P < 0.05 level, as determined by two-way analysis of variance. After exposure of calmodulin to anesthetic for 15 min followed by evaporation for 1 h to remove the anesthetic, the rightward shift was completely reversed, and Calcium²+ binding reverted to control values. At higher concentrations of isoflurane (8.7%) and halothane (5.7%), the leftward shift of the Calcium²+ -calmodulin binding curve induced by the anesthetics also was statistically significantly different from control. The leftward shift persisted even after the concentration of anesthetic in solution was measured to be zero, both for high halothane and for high isoflurane concentrations (Figure 3).

Titrations also were carried out with CaM^c. The control titration resulted in a K^dof 1.76 plus/minus 0.66 x 10 sup -6 M and n = 1.81 plus/minus 0.095. The same effect was observed for CaM^cas for CaM^swhen exposed to halothane and isoflurane. Figure 4(A) demonstrates the shift to the right at 0.57% (0.25 mM) halothane and the marked shift to the left at 5.7% (2.5 mM) halothane. Furthermore, with evaporation of halothane (5.7%, 2.5 mM), there was no reversibility (Figure 4(B)). The response to isoflurane of CaM^cwas qualitatively similar to that observed with CaM^s(data not shown).

The effect of low and high concentrations of halothane on CaM sub c were compared by two-way analysis of variance. The data sets obtained from Calcium²+ titrations during exposure to 0.57% and 5.68% halothane were statistically significantly different from the control data set (P < 0.05).

We have shown, by using intrinsic tyrosine fluorescence emission data, that the volatile anesthetics halothane and isoflurane alter calmodulin's Calcium²+ -binding properties. At low concentrations of anesthetic these effects appear reversible, whereas at higher concentrations they are irreversible for both isoflurane and halothane.

Our observations are based on the phenomenon that calmodulin in the presence of Calcium²+ undergoes a conformational shift with a corresponding enhancement of fluorescence. [9]Several sources of evidence support this finding. Spectroscopy studies have shown that Calcium²+ induces structural changes in the third and fourth Calcium sup 2+ -binding domains of calmodulin, where the two tyrosine residues lie. [16]Proteolysis studies have corroborated these findings by noting that calmodulin's third domain undergoes folding in the presence of Calcium²+, rendering the protein resistant to trypsin digestion. [17]Moreover, pK^aanalysis has shown that the environment of tyrosine 138, which resides in calmodulin's fourth domain, is altered in the presence of Calcium²+. [18,19].

The conformational change that occurs in the presence of Calcium²+, in addition to increasing fluorescence emission, exposes hydrophobic regions in calmodulin. [20]These hydrophobic surfaces have been implicated as necessary for calmodulin's stimulatory effects on other enzymes. [10]The same hydrophobic residues are also thought to be the sites where a host of lipophilic molecules, including the phenothiazines, interact with calmodulin to block its enzyme-activating steps. [20-25]Perhaps by binding to these hydrophobic sites the lipophilic volatile anesthetics alter calmodulin's Calcium²+ -binding properties.

We have shown two distinct effects of the volatile anesthetics on Calcium²+ binding. First, at low concentrations of anesthetic, calmodulin bound Calcium²+ with decreased affinity in a reversible fashion. In the presence of halothane and isoflurane, calmodulin may adopt a conformation that reduces its ability to bind Calcium²+. A reduction in the amount of Calcium²+ -bound calmodulin implies that the amount of calmodulin in the Calcium²+ -modified conformation available to activate other enzymes is decreased. Thus, the ability of the volatile anesthetics to promote amnesia, analgesia, respiratory depression, and muscle relaxation in a reversible fashion may result in part from temporary inhibition of the calmodulin-activated enzymes that normally regulate neural, respiratory, and skeletal muscle systems. [1].

In contrast, the second observed effect of the volatile anesthetics on calmodulin occurred at higher concentrations of halothane and isoflurane, much greater than concentrations used clinically. We observed a dose-dependent increase in affinity for Calcium²+. It is conceivable that at concentrations greater than some critical level, the hydrophobic anesthetics alter calmodulin's folding so that Calcium²+ binding is favored. At high concentrations of both anesthetics, the change in Calcium²+ binding was not reversed, even after the anesthetic concentrations were measured to be zero by gas chromatography. These studies were performed in the dark to prevent the possibility that photoactivation of the anesthetics would lead to a covalent interaction.

One factor in our experimental system that may account for an irreversible alteration in calmodulin structure and Calcium²+ binding is pH change. We observed a decrease in pH in our experimental solutions from 7.00 to 6.75 over the course of Calcium²+ additions in all titration experiments. Three observations, however, make it doubtful that pH shifts explain our results. First, data from Haiech et al. suggest that the pH decrease we encountered is not sufficient to alter Calcium²+ binding. Only at pH less than 6.0 is the affinity of calmodulin for Calcium²+ significantly altered. [8]Second, as noted by Haiech et al., a pH decrease confers a concomitant decrease in Calcium²+ affinity. [8]At high concentrations of anesthetics, we saw an increase in Calcium²+ affinity despite a slight pH decrement. Third, although we did observe a decrease in Calcium²+ -binding affinity at low concentrations of halothane and isoflurane, this decrease cannot be ascribed to a pH change, because although the pH change was irreversible, the decrease in Calcium²+ affinity that we observed reversed as the anesthetic evaporated.

Titrations in the presence of anesthetics were achieved by the liquid addition of buffer saturated with halothane or isoflurane. An artifact that may lead to the appearance of "irreversibility" may result from contamination of anesthetic-saturated buffers with Calcium²+. We therefore estimated the amounts of contaminating Calcium²+ in the saturated buffers by using the Calcium²+ -fluorescent dyes Calcium sup 2+ green and fura 2. We found slight amounts of contaminating Calcium sup 2+ in the anesthetic-saturated buffers, but they were much less than the amounts that could explain the marked shift to the left seen at high halothane and isoflurane concentrations.

Our finding that the volatile anesthetics can alter the conformation of a specific protein are corroborated by related results from a study by Franks and Lieb. They demonstrated that halothane and other general anesthetics can inhibit soluble firefly luciferase by binding at a hydrophobic pocket in the luciferase molecule. [26]Taken together, these experiments lend further credence to the theory that general anesthetics can act by specific interactions with target protein molecules and that nonspecific associations with the lipid bilayer component of cell membranes may not be required.

Although we now report significant effects of the volatile anesthetics on calmodulin's ability to bind Calcium²+, Blanck et al. were unable to show similar effects of halothane on troponin C. [27]Troponin C and calmodulin, both members of the EF-hand family of Calcium sup 2+ -binding proteins, share as much as 76% homology. In light of the considerable similarity between these two proteins, it is reasonable to expect that the volatile anesthetics may affect both similarly. However, given the dose-dependent, biphasic nature of the effects of the volatile anesthetics on calmodulin, we suggest that a significant alteration in Calcium²+ binding was not observed in the case of troponin C because of the concentration of halothane used in that experiment. In the troponin experiment, Calcium²+ binding was measured in the presence of 1.0 and 0.9 mM halothane. These concentrations are between the lower and higher concentrations of anesthetic that alternately yield a decrease or an increase in Calcium²+ binding in the case of calmodulin and may be pertinent to the lack of an observed alteration on Calcium²+ binding to troponin C.

In summary, by measuring the Calcium²+ -dependent fluorescence emission of calmodulin, we found that low concentrations of halothane and isoflurane reversibly decreased calmodulin's Calcium²+ affinity. At higher anesthetic concentration, however, Calcium²+ binding was increased in an irreversible fashion both by halothane and by isoflurane.

The authors thank Arthur Silvers, Ph.D., for suggestions for the statistical analysis of the Calcium²+ -binding curves; Velicia White for expert technical assistance; and Damian Blanck for the calculations. The authors also thank Olaf S. Andersen, M.D., for helpful discussions regarding protein conformation.

* Personal communication.