Cytochrome c oxidase inhibition by anesthetics: Thermodynamic ...

Proc. Nati Acad. Sci. USA Vol. 79, pp. 3749-3753, June 1982

Biophysics

Cytochrome c oxidase inhibition by anesthetics: Thermodynamic analysis (n-butanol/tetracaine/dibucaine/Arrhenius plot/luciferase)

GARRET VANDERKOOI AND BRAD CHAZOTTE Department of Chemistry, Northern Illinois University, DeKalb, Illinois 60115

Communicated by David E. Green, March 12, 1982

ABSTRACT The thermodynamic parameters that characterize the inhibition of cytochrome c oxidase activity, in rat liver submitochondrial particles, by n-butanol, tetracaine, and dibucaine were obtained. Three equilibria were assumed in order to account for the data: for the interaction of inhibitor with the native state of the enzyme, for the interaction of inhibitor with the thermally (reversibly) denatured state, and for the change between the native and thermally denatured states. Inhibition results from interaction with both the native and denatured states but, because the interaction is stronger with the denatured than with the native state, the native/denatured equilibrium is shifted to the right by the anesthetics. The enthalpies of interaction are -2.3, -4.7, and 3.7 kcal/mol (1 cal = 4.18 J) for the native state and -10, -6, and -14 kcal/mol for the denatured state, for n-butanol, tetracaine, and dibucaine, respectively. These values are much smaller than the previous estimates obtained by using the assumption that anesthetics interact only with the thermally denatured state of enzymes (e.g., -81 kcal/mol for tetracaine inhibition of luciferase). Our results suggest that local anesthetics inhibit enzyme activity by causing a reversible perturbation of protein conformation. The magnitude of the perturbation is much smaller (in energetic terms) than that which accompanies thermal denaturation.

the same value being given as positive or negative in different places. The correct sign for AH3 in the studies referred to in refs. 4-6 is negative-i. e., the binding reaction is exothermic. However, Eyring et al. (7) considered AH3 to be positive (i.e., the same sign as for denaturation), and their proposal that anesthetics cause protein unfolding was based on this assumption. We have used the general methods of Johnson et al. (1, 2) to interpret our data on the inhibition of cytochrome c oxidase activity by anesthetics (9). We found, however, that it was impossible to account simultaneously for the inhibition data in the high- and low-temperature ranges by using the assumptions of either type I or type II inhibition defined by Johnson et al. (1). (In type I the inhibitor binds indiscriminately to the native and denatured states whereas in type II the inhibitor binds only to the denatured state.) An excellent fit to the data was obtained by assuming that the anesthetics interact with both the native and denatured states but do so unequally. The AH values that we obtain by using this more general approach are much smaller than the values previously reported (4-6), and thus our results do not support the proposal of Eyring et al. (7) that a major unfolding of protein structure accompanies anesthetic inhibition.

Johnson et al. (1, 2) showed how it is possible to obtain thermodynamic information on inhibitor binding to proteins by studying the temperature dependence of enzyme activity. The information thus obtained can reveal whether the inhibitor binds preferentially to the native or denatured forms of the enzyme, and the magnitude of the enthalpy and entropy changes may indicate whether any appreciable conformational alterations accompany the binding. Johnson et al. (1-4) carried out thermodynamic analyses of the inhibition of bacterial luminescence and respiration by anesthetics and other substances. More recently, the same approach was used by Ueda et al. (5, 6) to study the inhibition of purified firefly luciferase by general and local anesthetics. In each of these studies, large negative enthalpy values were found for the binding of anesthetics to the denatured form ofthe enzyme [e. g., -81 kcal/mol for luciferase inhibition by tetracaine (6)]. These values were of the same order of magnitude as, but of opposite sign to, the enthalpies of denaturation reported for the same systems. On this basis, Eyring et al. (7) proposed that anesthetics cause a major unfolding of protein structure, comparable in extent to protein denaturation. The well known pressure reversal of anesthesia (2, 8) was also explained in terms of this unfolding: if a volume increase accompanies the unfolding, pressure should counteract the anesthetic effect. There has been some confusion in the literature over the sign of the enthalpy change accompanying anesthetic binding, with

THERMODYNAMIC METHODS The following scheme is used for the thermodynamic analysis of inhibition (1, 2): K1

En

z

Ed

+

+

rU

sU

[1]

K3

K2

EnUr

EnUs

in which En is the native, uncomplexed enzyme and is assumed to be the only active form; Ed is the thermally (reversibly) denatured enzyme; U is the inhibitor substance; and r and s are the apparent numbers of molecules of U that bind to the native and denatured enzyme forms in a cooperative manner. The Ks are the equilibrium constants, which are related to the respective AH and AS values in the usual fashion: ln K = -AH/RT + AS/R. The enzyme activity is proportional to the amount of enzyme in the En state. In the absence of inhibitor, this is given by:

VO = 1-kEo + K1

[2]

in which Eo is the total of all enzyme species, and k is the rate constant. The temperature dependence ofthe rate constant may be expressed by absolute rate theory in terms of AHt, the activation enthalpy: k = cT exp(-AHt/RT). The entropy of ac-

The publication costs ofthis article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U. S. C. §1734 solely to indicate this fact.

3749

3750

Biophysics: Vanderkooi and Chazotte

Proc. Natl. Acad. Sci. USA 79 (1982)

tivation is included in c, which is a collection of constants. The rate in the presence of inhibitor is given by VI: kEo [3] VI = 1 + K2Ur + K1 (1 + K3Us)

Johnson et al. (1) showed how to plot experimental inhibition data in linearized form for two special cases ofEq. 1 (type I and type II inhibition) in order to obtain the relevant thermodynamic parameters. We show here how to treat the general case in which the inhibitor binds unequally to the native and denatured states and also show how the two special cases are related to the general case. Johnson et al. (2) defined a rF function as follows: [4]

Fl = - 1.

VI

From the rate expressions for VO and V1 (Eqs. 2 and 3), the general theoretical expression for F1 based on the scheme in Eq. 1 is:

K2ur + K K3Us [5] 1 + K, This equation reduces to the F1 function given by Johnson et al. (2) for the special case of type I inhibition, in which- K2 = K3 and r = s. [6] F, (type I) = K2Ur .

The latter expression is also obtained from Eq. 5 in the low temperature limit, where K1 1 (see Table 1). (Hyperbolic curves would be expected if r and s were equal to 1; this can be readily demonstrated for the low-temperature limit by solving Eq. 6 for V/Vo..) The nonintegral values obtained for r and s are an indication that there is less than complete cooperativity among the several ligand molecules required to give inhibition; the actual number of molecules of inhibitor that interact must be greater than the apparent r and s values.

DISCUSSION A fundamental question is whether it is appropriate to analyze the inhibition of a membrane enzyme by using the approach of Johnson et al. (1), which is described in terms of ligand binding to proteins, when the ligands involved are also known to be lipid perturbants. A negative answer probably must be given if the observed inhibition is the result of a generalized nonspecific fluidization of the membrane lipid. If, on the other hand, the anesthetics bind directly to the protein at the membrane-water interface, or else interact laterally by dissolving in the boundary lipid (lipid molecules that are required for enzyme activity and hence may be called part of the holoenzyme), the present method of analysis should be valid. Although one cannot absolutely rule out a lipid perturbation mechanism for cytochrome c oxidase inhibition (the membrane counterpart of nonspecific solvent effects for soluble proteins), it appears unlikely that the

X 50

Dibucaine, mM FIG. 3. Relative rate plotted as a function of dibucaine concentration for various temperatures: *, 2000; 0, 3500; A, 39-C; A, 430C.

3752

Biophysics: Vanderkooi and Chazotte


detailed differences between tetracaine and dibucaine inhibition could be accounted for in this manner. The temperature coefficient of inhibition, for example, is negative for dibucaine but is positive for tetracaine (as well as for n-butanol) in the lowtemperature range; this difference is reflected in the opposite signs for AH2 (Fig. 2; Table 1). This indicates that the molecules responsible for the enzyme inhibition must approach the protein closely enough to be "seen" by the protein, because no similar degree of discrimination between these structurally related molecules is to be expected from the lipid. Direct interaction of local anesthetics with proteins is not without precedent; we have shown that water-soluble F1-ATPase is inhibited by these agents, with the inhibitory potency being proportional to the octanol/water partition coefficients (13, 14). One might well ask why we obtain small AH values for cytochrome c oxidase inhibition, whereas much larger values were reported for luciferase (5, 6) and other systems (2, 4). It appears that the difference is not simply due to the fact that different enzyme activities were assayed but rather is due to the method of analysis used. The reason will become apparent through consideration of Fig. 4, which shows plots of the various r functions vs. 1/T. The data given are for the case of cytochrome c oxidase inhibition by 290 mM n-butanol. The theoretical curve for In rF was calculated by using Eq. 5 and the thermodynamic parameters given in Table 1. It is evident that this curve fits the experimental points over the entire temperature range (15-45TC). According to Eq. 6, the slope of the linear portion at low temperature equals -AH2/R. (The values of AH2 given in Table 1 were determined from the average of the slopes of plots ofthis type for all the anesthetic concentrations used.) Fig. 4 also shows plots of In F2 and In r3 vs. 1/T. That for In r3 is linear, but that for In '2 consists of two linear portions connected by a curve. (In computing the theoretical In F2 curve with Eq. 8, Eq. 5 was used to calculate F1.) At high temperature, In r2 asymptotically approaches In r3 as expected from Eqs. 7 and 8. The slope of In F3, as well as that of the high-temperature limit of in F2, equals -AHJIR. (Experimentally, only the high-temperature data can be used with either In F2 or In F3 to determine AH3 because, at lower temperatures, small errors in the data are greatly magnified by the calculations.) 16

I

+1 4e.

12 p

/-+ 8

+ , ..~~~~~~~4..

m

.'IgS -6;~1-

\--~~~~~~~~.

4 .0

m~

0~4~

0 -

.

6-

0

0

I

3.1

I

3.2

I

I

3.3

1/T, K

x

3.4

3.5

103

FIG. 4. Logarithmic plot of the F functions vs. 1/T for the case of 290 mM n-butanol inhibition of cytochrome c oxidase. See text for details. * and -, ln F7; + and -. -., ln F2; o and ---, ln r13.

The slope of the low-temperature arm of the In r2 curve is approximately AH1/R. This is the case because In r2 = In rF + ln(l + 1/K1); at low temperature the slope of In r1 is small (being equal to -AH2/R) and hence the slope of In r2 is dominated by the slope of the second term, which is AH1/R. In the case of luciferase inhibition by tetracaine (6), the published graphs of In rl and in F2 vs. 11T are essentially linear over the entire temperature range reported (figure 5 of ref. 6). According to Johnson et al. (2), a linear In r1 plot implies type I inhibition, whereas a linear In r2 plot implies type II inhibition. Ueda et al. (6) therefore concluded that the inhibition was a "hybrid of type I and type II." They determined AH3 to be -81 kcal/mol (incorrectly reported as +81 kcal/mol), from the slope of In r2. This value may be compared with the heat of denaturation of luciferase (AH1), which was 79 kcal/mol. To us it appears that the luciferase system behaves similarly to cytochrome c oxidase and hence should be analyzed by the general scheme in Eq. 1. If this is so, one must conclude that the data given in figure 5 of ref. 6 cover only the low-temperature arms ofthe In r1 and In r2 functions (i.e., the portions corresponding to 1T > 0.0033 in Fig. 4 here). Because the slope of the lowtemperature arm of In F1 equals -AH2/R, we estimate from the published figure that AH2 for luciferase inhibition by tetracaine is about -5.5 kcal/mol. The slope ofthe low-temperature arm of In F2 is approximately AH1/R, however, and not -A&HJR (as interpreted). No estimate can be made of AH3 for luciferase from the published data because the data were not reported at sufficiently high temperatures. We conclude from our studies that the alcohol and tertiary amine anesthetics interact with both the native and denatured states of cytochrome c oxidase to cause inhibition. Whether this interaction is with the protein per se or with the boundary lipid intimately associated with the protein cannot be determined from the data in hand. A qualitatively similar type of inhibition was found with several other segments of the mitochondrial electron transport chain (9, 12) and, by our interpretation, the inhibition of luciferase follows the same pattern. Our finding that AH2 and AH3 for anesthetic inhibition are small in comparison with the heat of denaturation casts doubt on the theory of Eyring et al. (6, 7) for the mechanism of action of anesthetics, which was based on the assumption that anesthetic inhibition is accompanied by a large, positive heat change. Nevertheless, some qualitative features of their model may still be valid. Because we find that the anesthetics shift the native/ denatured equilibrium to the right, by interacting more strongly with the denatured than with the native form, a significant conformational change equivalent to denaturation may be induced by the action of anesthetics. This effect would only be expected at the higher temperatures at which the denaturation equilibrium is active. At lower temperatures, the conformational changes presumed to be responsible for the observed enzyme inhibition must be of a much smaller magnitude (in energetic terms) than those involved in thermal denaturation. Hence, the tertiary amine local anesthetics may constitute a class of heretofore unrecognized relatively weak, reversible perturbants of protein tertiary or quaternary structure. We thank Dr. J. Erman for helpful discussions on the thermodynamic analysis in this paper. This work was supported in part by grants from the American Heart Association and by National Institutes of Health Biomedical Research Support Grant RR 07176. 1. Johnson, F. H., Eyring, H. & Williams, R. W. (1942) J. CelL Comp. PhysioL 20, 247-268. 2. Johnson, F. H., Eyring, H. & Polissar, M. J. (1954) The Kinetic Basis of Molecular Biology (John Wiley, New York), pp. 369-514. 3. Johnson, F. H., Eyring, H., Steblay, R., Chaplin, H., Huber, C. & Gherardi, G. (1945)J. Gen. Physiot 28, 463-537.

Biophysics: Vanderkooi and Chazotte 4. Koffler, H., Johnson, F. H. & Wilson, P. W. (1947) J. Am. Chem. Soc. 69, 1113-1117. 5. Ueda, I. & Kamaya, H. (1973) Anesthesiology 38, 425-436. 6. Ueda, I., Kamaya, H. & Eyring, H. (1976) Proc. Natl. Acad. Sci. USA 73, 481-485. 7. Eyring, H., Woodbury, J. W. & D'Arrigo, J. S. (1973) Anesthesiology 38, 415-424. 8. Johnson, F. H. & Flager, E. A. (1951) Science 112, 91-92. 9. Chazotte, B. & Vanderkooi, G. (1981) Biochim. Biophys. Acta 636, 153-161.


3753

10. Lenaz, G., Sechi, A. M., Parenti-Castelli, G., Landi, L. & Bertoli, E. (1972) Biochem. Biophys. Res. Commun. 49, 536-542. 11. Lee, M. P. & Gear, R. L. (1974)J. BioL Chem. 249, 7541-7549. 12. Vanderkooi, G., Chazotte, B. & Biethman, R. (1978) FEBS Lett. 90, 21-23. 13. Vanderkooi, G., Shaw, J., Storms, C., Vennerstrom, R. & Chignell, D. (1981) Biochim. Biophys. Acta 635, 200-203. 14. Chazotte, B., Vanderkooi, G. & Chignell, D. (1982) Biochim. Biophys. Acta, in press.