How Do Volatile Anesthetics Inhibit Ca 2 (cid:49) -ATPases?*

Volatile anesthetics at concentrations that are used in clinical practice to induce anesthesia selectively inhibit activity of the plasma membrane Ca 2 (cid:49) -transport ATPase (Kosk-Kosicka, D., and Roszczynska, G. (1993) Anesthe- siology 79, 774–780). We have investigated the mechanism of the inhibitory action of several anesthetics on the purified erythrocyte Ca 2 (cid:49) -ATPase by employing fluorescence spectroscopy measurements that report changes in the environment of intrinsic tryptophans and of an extrinsic probe attached in the active site of the enzyme. We have shown that the observed inhibition of the Ca 2 (cid:49) -dependent activation of the enzyme correlates well with the elimination of the Ca 2 (cid:49) -induced conformational change that is important for the proper function of the enzyme. Analysis of the anesthetics effects on the total tryptophan fluorescence indicates a significant effect on enzyme conformation. Similar changes have been observed in the sarcoplasmic reticu- lum Ca 2 (cid:49) -ATPase. We propose that volatile anesthetics inhibit Ca 2 (cid:49) -ATPase by interacting with nonpolar sites in protein interior, in analogy to the binding demon- strated for myoglobin, hemoglobin, and adenylate ki-nase (Schoenborn, Featherstone, Such binding is expected to modify conformational substate(s) The concentrations of volatile anesthetics in the in vitro assays described in our publication are 10-fold higher than indicated. As a result, only the initial activation of the SR Ca 2 (cid:49) -ATPase, in contrast to the inhibition of the plasma membrane Ca 2 (cid:49) -ATPase, occurs at clinical concentrations of volatile anesthetics. The K app calculated from tryptophan quenching experiments using the corrected anesthetic concentrations are about 10-fold lower, i.e. more comparable to the values obtained for other proteins. The concentration error does not affect the main findings of our paper which suggest that the anesthetics interfere with enzymatic function by binding in nonpolar sites in protein interior and modifying conformational substate(s) of the enzyme.

The molecular site and mechanism of action of inhaled volatile anesthetics is not known despite intensive studies of over 30 years. Various hypothesis have been developed ranging from lipids to proteins as molecular targets and unspecific to specific sites of action (for review, see Miller (1985), Koblin (1990), and Franks and Lieb (1994)). In recent years the focus has shifted to membrane proteins, including ion channels and receptors for which different anesthetic effects have been reported, however, neither a specific nor an unifying mechanism of action has been described. We have previously identified another intrinsic membrane protein, the plasma membrane Ca 2ϩ pump, as a potential target for anesthetic action (Kosk-Kosicka and Roszczynska, 1993;Kosk-Kosicka, 1994). Its Ca 2ϩ -ATPase activity and Ca 2ϩ transport are inhibited by a variety of inhaled anesthetics, both clinically used and experimental ones (Kosk-Kosicka and Roszczynska, 1993;Franks et al., 1995). 1 We have described the effects of four halogenated volatile anesthetics that are used in clinical anesthesia on the normal process of enzyme activation by either calmodulin binding or dimerization (Kosk-Kosicka and Roszczynska, 1993). All four drugs inhibited these two activation pathways in a dose-dependent manner and with a similar I 50 . The inhibition was observed at their clinical concentrations suggesting that Ca 2ϩ -ATPase was not only a good model target of an intrinsic membrane protein but could also be a pharmacological in vivo target for this group of general anesthetics. In subsequent studies we have demonstrated that the inhibition was selective as judged by the following criteria. First, two other groups of general anesthetics, barbiturates and alkanols, did not inhibit the enzyme at their anesthetic concentrations (Kosk-Kosicka et al., 1992). 1 Second, several other ATPases including Mg 2ϩ -ATPase and Na ϩ ,K ϩ -ATPase showed significantly lower sensitivity to the volatile anesthetics than did the Ca 2ϩ -ATPase. 2 Third, the above phenomena were observed in three distinct plasma membranes, ranging from erythrocytes to neuronal and endothelial cell types (Kosk-Kosicka et al., 1995b). As the site and mechanism of anesthetic action have not been explained and only a few proteins among the several reported to be targeted by anesthetics in vitro are affected at clinically relevant concentrations the observed inhibition of the Ca 2ϩ -ATPase activity deserves a close attention.
In the present study we have investigated the mechanism of enzyme inhibition by volatile anesthetics by two experimental approaches: Ca 2ϩ -ATPase activity assay and fluorescence spectroscopy measurements. We have used dimeric Ca 2ϩ -ATPase whose activation is independent of a modulatory protein calmodulin since we have previously established that both normal modes of enzyme activation (by calmodulin binding to enzyme monomers and by the self-association of monomers to dimers) are equally sensitive to this group of general anesthetics (Kosk-Kosicka and Roszczynska, 1993;Kosk-Kosicka and Bzdega, 1988;. We have assessed the effects of the anesthetics on the Ca 2ϩ -dependent conformational changes of the enzyme by monitoring the Ca 2ϩ -dependent changes in fluorescence intensity of two probes: 1) intrinsic tryptophan(s) that reflect a conformational change which the Ca 2ϩ -ATPase undergoes upon binding the substrate Ca 2ϩ in the initial step of enzymatic cycle, and 2) an external probe, fluorescein 5Ј-isothiocyanate (FITC) 3 attached to lysine 601 in the active site that normally binds ATP (Dupont, 1976;Inesi et al., 1980;Kosk-Kosicka and Inesi, 1985;Kosk-Kosicka et al., 1989). We demonstrate that the anesthetics inhibit both measures in a dose-dependent manner and there is a good correla-tion between the attenuation of the Ca 2ϩ -dependent conformational changes and the inhibition of the Ca 2ϩ -ATPase activity. In addition, the observed changes in the total tryptophan fluorescence also suggest that the anesthetics affect enzyme conformation. We analyze these findings with respect to the demonstrated binding of small ligands such as xenon (a very potent experimental inhaled anesthetic) in interior spaces of metmyoglobin which affects the internal motions and substates of the protein (Schoenborn and Featherstone, 1967;Tilton et al., 1984). As proteins undergo motions small molecules enter and interact with nonpolar sites in the protein interior (Englander et al., 1972;Lakowicz and Weber, 1973;Eftink and Ghiron, 1976;Cooper, 1976;Cohen et al., 1977;Lim and Sauer, 1991;Eriksson et al., 1992;Lim et al., 1994). We postulate that interaction of anesthetic molecules with nonpolar sites in the Ca 2ϩ -ATPase molecule modifies conformational substate(s) of the protein which results in impairment of its enzymatic function. We consider binding of small molecules in nonpolar internal protein spaces a general phenomena whose occurrence, however, requires compatibility between the nonpolar sites available in the protein and the invading molecule. In the case of gaseous anesthetic such an interaction may or may not have functional consequences at clinical anesthetic concentrations depending on the structure (flexibility) and function of a given protein, as it does for the Ca 2ϩ -ATPases but apparently not for myoglobin or hemoglobin. The lack of a significant functional effect of the demonstrated anesthetic binding in myoglobin or hemoglobin and their relationship to anesthesia apparently eliminated it from consideration as a mechanism of anesthetic action. In contrast, both function and conformation of the Ca 2ϩ -ATPase are significantly disturbed by the anesthetics and their action on this intrinsic membrane protein which controls intracellular Ca 2ϩ homeostasis could certainly contribute to the pharmacological anesthetic effects. Our model needs to be treated as hypothetical until future advances in NMR or x-ray crystallography allow for its verification.

MATERIALS AND METHODS
Egg yolk phosphatidylcholine (P5763) and CNBr-activated Sepharose 4B were purchased from Sigma; octaethylene glycol mono-n-dodecylether (C 12 E 8 ) was obtained from Nikko (Tokyo, Japan). Coupling of bovine calmodulin to Sepharose was performed in accordance with Pharmacia Biotech Inc. instructions as described earlier (Kosk-Kosicka and Bzdega, 1988). Enflurane and isoflurane were obtained from Anaquest (Liberty Corner, NJ); methoxyflurane was from Abbott Hospital Products; and thymol-free halothane from Halocarbon Laboratories (River Edge, NJ). Halothane is a two-carbon alkane halogenated derivative, the other three anesthetics belong to halogenated methyl ethyl ether derivatives as shown in Fig. 1.
The methods used for preparation of erythrocyte ghost membranes and sarcoplasmic reticulum (SR), purification of the Ca 2ϩ -ATPase from erythrocyte membranes, determination of protein, and Ca 2ϩ concentration were as described previously (Kosk-Kosicka et al., 1983Kosk-Kosicka and Bzdega, 1988). Free Ca 2ϩ concentrations were calculated (Fabiato and Fabiato, 1979) from total calcium and EGTA concentrations, based on the constants given by Schwartzenbach et al. (1957). Total calcium was measured by atomic absorption. Sarcoplasmic reticulum was prepared from rabbit skeletal muscle in the laboratory of Dr. Inesi (Eletr and Inesi, 1972).
The Ca 2ϩ -ATPase Activity-This was determined by colorimetric measurement of inorganic phosphate production, generally as described previously (Kosk-Kosicka and Bzdega, 1988;Kosk-Kosicka et al., 1983). The assay was performed in sealed (to maintain constant concentration of the anesthetic) 0.65-ml polypropylene tubes in a total reaction volume of 0.33 ml. After addition of all reagents and protein, volatile anesthetic was delivered to the tube in an air-tight Hamilton syringe and the reaction was started with 3 mM ATP. After vortexing, samples were placed in a water bath at either 37 or 25°C for 30 min. The reaction was terminated with ammonium molybdate/metavanadate at individual times. The aliquots of volatile anesthetics delivered to the assay were taken from solutions of saturated volatile anesthetics in reaction mixture, which were prepared daily from the stock of pure volatile anesthetic under nitrogen gas at room temperature. Activity of the purified dimeric Ca 2ϩ -transport ATPase was assayed at 70 nM enzyme concentration in a reaction mixture containing 50 mM Tris maleate, pH 7.4, 120 mM KCl, 8 mM MgCl 2 , 3 mM ATP, 1 mM EGTA, and CaCl 2 in concentrations yielding the required free Ca 2ϩ . The concentration of C 12 E 8 was kept constant at 150 M. For activity assay of the Ca 2ϩ -transport ATPase in membranes, C 12 E 8 was omitted. The activity was calculated as a difference between the activity determined in the above reaction mixture and the reaction mixture in which Ca 2ϩ was omitted. 65 g of ghost membranes and 1.2 g of sarcoplasmic reticulum were used per tube. In addition, the reaction mixture for sarcoplasmic reticulum contained the divalent cation ionophore A23187 (10 M).
Fluorescence Measurements-Fluorescence intensity was measured at equilibrium using a FluoroMax spectrofluorometer with DM3000F software. Tryptophan fluorescence was excited at 290 nm and the emission was recorded at 330 nm. Excitation and emission wavelengths used for FITC Ca 2ϩ -ATPase were 484 and 515 nm. The Ca 2ϩ -ATPase in erythrocyte ghosts was labeled with FITC as described previously . The labeled enzyme was then purified by our standard procedure. The stoichiometry of labeling was 1 mol of FITC/1 mol of enzyme. Fluorescence measurements were done in the medium containing 100 mM Tris-HCl, pH 7.4, 150 M C 12 E 8 , 130 mM KCl, 8 mM MgCl 2 , and 1 mM EGTA at 25°C in the total volume of 1.1 ml in a sealed cuvette. Protein concentration was 10 g/ml for the purified plasma membrane Ca 2ϩ -ATPase (70 nM) and SR. For measurements of the calcium-dependent change in fluorescence CaCl 2 was added after the anesthetic and the calcium-induced decrease of FITC fluorescence was related to the total initial fluorescence at each anesthetic concentration (F i ). Because total tryptophan fluorescence decreased with time we have differentiated between F io (at time 0) and F i (measured right before the addition of CaCl 2 ). Free Ca 2ϩ was 17.5 M. The solution was gently stirred during the experiment which lasted 4 -5 min. Stirring did not cause any perturbation of the fluorescence signal. Corrections for the dilution effect were made when necessary.
Decays of Fluorescence Intensity-The frequency domain data were analyzed by the method of nonlinear least squares Lakowicz et al., 1984). The measured values were compared with values predicted from a model. For a multiexponential decay, I(t) is given by, where ␣ i is the pre-exponential factor and i is the lifetime. The fractional intensity of each component in the decay is given by:

FIG. 1. Chemical structures of volatile anesthetics used in the study.
the best fit between the data and the calculated value is indicated by a minimum value for the goodness-of-fit parameter 2 R : where is the number of degrees of freedom and ␦ and m are uncertainties in the phase and modulation values, respectively. Determination of Concentrations of Anesthetics-In parallel to the activity or fluorescence assay, samples were incubated under identical conditions and used for determination of the effective anesthetic concentrations during the assay. Halothane, isoflurane, and enflurane were extracted with heptane (methoxyflurane with octane), measured by gas chromatography, and the concentrations were calculated as described earlier (Kosk-Kosicka and Roszczynska, 1993). 1 MAC equals minimum alveolar concentration of an agent that produces immobility in 50% of those subjects exposed to a noxious stimulus as defined by Eger (1978).
Data are expressed as the mean Ϯ S.E. of three to six independent experiments performed in duplicates. The data points were fitted by the nonlinear regression method. RESULTS We have demonstrated a strong correlation between the general anesthetic potency (as expressed in MAC values) of several halogenated volatile anesthetics and their ability to inhibit the Ca 2ϩ -ATPase activity (Kosk-Kosicka and Roszczynska, 1993). The anesthetic concentrations in our activity assays were expressed in volume % at 37°C for easy comparison with their anesthetic potency. At present we have recalculated the values into millimolar concentrations (in addition, original data for methoxyflurane have been included) and compared the concentrations of the anesthetics in the reaction mixture that halfmaximally inhibit the purified erythrocyte Ca 2ϩ -ATPase with their concentrations in blood to which the enzyme may be exposed when 1 MAC of a given anesthetic is administered in clinical anesthesia. Fig. 2 shows that the concentrations of all four agents that the enzyme may encounter during anesthesia inhibit enzyme function.
The anesthetics decrease both V max and Ca 2ϩ affinity of the Ca 2ϩ -ATPase activity (Fig. 3). As shown in Fig. 3 isoflurane at a concentration that half-maximally inhibits the enzyme causes a small yet statistically significant shift in the apparent K1 ⁄2 for calcium from pCa 7.05 Ϯ 0.05 to 6.9 Ϯ 0.04. Similar results were obtained with other anesthetics (data not shown).
The effects of volatile anesthetics on the enzyme were subsequently examined at 25°C by two experimental approaches: Ca 2ϩ -ATPase activity assay and fluorescence spectroscopy measurements. Fig. 4 shows that at this temperature the activity is also inhibited in a dose-dependent manner, similar to the patterns observed at 37°C except that higher concentrations are required for a comparable extent of inhibition. This temperature dependence is in agreement with nonpolar interactions between each anesthetic and the enzyme (Collins and Washabaugh, 1985;Makhatadze and Privalov, 1995).
The effect of anesthetics on the functionally important enzyme conformation has been assessed by measuring the calcium-dependent increase in intrinsic tryptophan fluorescence induced by substrate Ca 2ϩ binding and in addition, by the decrease of the fluorescence of an external probe, FITC, bound in the active site of the enzyme. As shown in Fig. 5 the Ca 2ϩinduced increase in tryptophan fluorescence is attenuated by all three anesthetics in a similar dose-dependent manner. Halothane effect could not be determined accurately due to the high extent of quenching of tryptophan fluorescence by this compound. The half-maximal reduction is observed at ϳ0.16 -0.26 mM concentrations of the agents in the order of efficiency shown in Fig. 5; the difference between their F 50 values may be not statistically significant. The Ca 2ϩ -induced fluorescence increase is readily reversed by addition of EGTA both in the presence and absence of anesthetics. Similar dose dependence is observed in the attenuation of the Ca 2ϩ -dependent decrease in FITC fluorescence, even though the ⌬F/⌬F Total does not proceed to 0 (Fig. 6). Thus, there is a good correlation between the inhibitory action of the anesthetics on the Ca 2ϩ -ATPase activity (I 50 ϭ 0.20 -0.27 mM) and the Ca 2ϩ -induced conformational change assessed by either tryptophan or FITC fluorescence measurements (F 50 ϭ 0.16 -0.23 mM).
We have also analyzed the anesthetics effects on the total tryptophan fluorescence intensity of Ca 2ϩ -ATPase. All four studied compounds decrease the fluorescence in a dose-dependent linear fashion (Fig. 7). In order to distinguish which type of quenching process is taking place, lifetime fluorescence inten- FIG. 2. Correlation between anesthetic and inhibitory potencies of four inhalation anesthetics: isoflurane (q), methoxyflurane ( ), enflurane (f), and halothane (*) at 37°C. The anesthetic potencies, expressed in millimolar concentrations, in blood were calculated from published values of MAC taking into consideration the appropriate blood/gas partition coefficients (Eger, 1974;Koblin, 1990;Kosk-Kosicka and Roszczynska, 1993). Inhibitory potencies are expressed in millimolar concentrations (in the reaction mixture) that half-maximally inhibit Ca 2ϩ -ATPase activity of the purified dimeric enzyme. Both sets of values were obtained at 37°C, i.e. human body temperature.
FIG. 3. Calcium dependence of Ca 2؉ -ATPase activity in the presence (q) and absence of isoflurane (E). The Ca 2ϩ -ATPase activity was assayed as described under "Materials and Methods." The reaction mixture contained 50 mM Tris maleate, pH 7.4, 120 mM KCl, 8 mM MgCl 2 , 150 M C 12 E 8 , 1 mM EGTA, and 3 mM ATP. Various amounts of CaCl 2 were added to obtain the calcium concentrations specified in the horizontal axis. Enzyme concentration was 70 nM. The assays were performed at 37°C. Inset shows normalized data of one typical experiment, where 100% is the maximal activity either with or without isoflurane. Six independent experiments were performed in duplicates. The mean K1 ⁄2 for the enzyme in the presence of isoflurane was significantly different from the control (p Ͻ 0.04). sity measurements were performed in the presence and absence of isoflurane. As shown in Table I intensity decays are very heterogenous and three exponentials are needed to fit the data. This is expected for a protein with several tryptophans and was observed previously for the SR Ca 2ϩ -ATPase (Lakowicz et al., 1986;Gryczynski et al., 1989;Wang et al., 1992). In the presence of isoflurane the lifetimes decrease (compare exponential fits 2 R and ϽϾ in Table I). These changes in lifetimes are characteristic of collisional quenching (Pesce et al., 1971;Lakowicz, 1991). On the other hand the apparent quenching constants, K app , calculated from the decrease in fluorescence intensity (in Fig. 7 and Table II) are larger than those possible if collisional quenching were entirely responsible for the observed fluorescence decrease (for example, for other multitryptophan proteins ␤-trypsin and pepsin the K app are 2.4 M Ϫ1 and 9.5 M Ϫ1 ), indicating contributions from other processes such as binding of the quencher and conformational changes (Sellers and Ghiron, 1973;Eftink and Ghiron, 1976;Lakowicz, 1991). Accordingly, steady-state fluorescence intensity spectra recorded for the enzyme in the absence and presence of increasing concentrations of isoflurane show a progressive decrease of intensity and a red shift of up to 6 nm at 0.52 mM isoflurane concentration (Fig. 8), suggesting an increased polarity of the tryptophan environment that could result from a conformational change caused by the interacting anesthetic. It has been shown that conformational changes caused by partial denaturation of protein result in fluorescence intensity and lifetime decreases and a red shift of the spectrum (Gryczynski et al., 1988). To assess the contribution of a direct quenching of tryp- tophans by the anesthetics we have tested the effect of anesthetics on the fluorescence of the heat-denaturated enzyme and on free tryptophan in the same reaction mixture, including detergent. The free tryptophan fluorescence was not affected while there was a decrease in the fluorescence of the denatured enzyme. In addition, no fluorescence resonance energy transfer was observed between tryptophan and anesthetics (no overlap of absorption and emission spectra). All these findings point to a substantial contribution of conformational change of the enzyme in the anesthetic effect on the total tryptophan fluorescence.
To examine the contribution of lipids in the decrease of total tryptophan fluorescence due to anesthetics we investigated their action on the sarcoplasmic reticulum Ca 2ϩ -ATPase. This enzyme could be studied in the native membrane because it constitutes Ϸ80% of the total membrane protein as compared to Ϸ0.01% of the membrane protein represented by the Ca 2ϩ -ATPase in the plasma membrane. Tryptophan fluorescence of the SR enzyme is reduced by the four anesthetics with the same order of efficiency as observed for the purified plasma membrane Ca 2ϩ -ATPase (Table II). Also similar to the plasma membrane enzyme, the apparent quenching constants are too large for an exclusively diffusion controlled process. They are approximately two times lower than for the purified plasma membrane enzyme. The difference could arise from either an interference of the surrounding lipids or differences in the enzyme protein molecules such as existence of fewer sites for interaction with the anesthetics in the SR enzyme or their lower affinity than in the plasma membrane enzyme. The fact that about 2-fold lower anesthetic concentrations are required for inhibition of the purified plasma membrane Ca 2ϩ -ATPase activity as compared to the enzyme in the erythrocyte ghost membrane ( Fig. 9: I 50 ϭ 0.26 and 0.50 mM halothane) suggests that the lower apparent quenching constants for the enzyme in SR may be caused by a decrease of tryptophans accessibility to the anesthetics by the lipid bilayer. A similar difference in halothane sensitivity of the Ca 2ϩ -ATPase in ghost membrane versus purified was observed previously at 37°C (Kosk-Kosicka and Roszczynska, 1993). As shown in Fig. 9, the Ca 2ϩ -ATPase in SR requires a 40 -50% higher halothane concentration for half-maximal inhibition than the enzyme in erythrocyte membrane. This result and the fact that at concentrations up to 0.2 mM halothane the enzyme in SR is activated, in contrast to the plasma membrane Ca 2ϩ -ATPase (Fig. 9), indicate that in addition to the lipid effect, the nonpolar sites in the SR enzyme may have a lower affinity for the anesthetics than comparable sites in the plasma membrane enzyme. DISCUSSION We have shown that volatile anesthetics affect both the conformational state and activity of the Ca 2ϩ -ATPase. There is a strong correlation between the attenuation of the Ca 2ϩ -induced functionally important conformational change of the protein and impairment of its enzymatic function by anesthetics. Also our data on anesthetic effects on the total tryptophan fluorescence, including the decrease of fluorescence intensity, the red shift, and the decrease in lifetimes, indicate that anesthetics produce conformational changes in the enzyme molecule. To explain these effects we suggest that the anesthetics enter the protein and interact with nonpolar amino acids in its interior. To accommodate the intruding anesthetic the protein molecule undergoes structural rearrangement that results in the loss of its Ca 2ϩ -ATPase activity. In proposing this mechanism of anesthetic action on the Ca 2ϩ -ATPase we are considering multiple x-ray diffraction and NMR data that have demonstrated binding of gaseous anesthetics and other molecules, including urea and steroids, in hydrophobic "cavities" in the interior of several proteins (Schoenborn, 1965;Schoenborn et al., 1965;Schoenborn and Featherstone, 1967;Nunes and Schoenborn, 1973;Brown et al., 1976;Sachsenheimer et al., 1977;Hibbard and Tulinsky, 1978;Tilton and Kuntz, 1982;Tilton et al., 1984;Otting et al., 1991;Arevalo et al., 1994;Jain et al., 1994;Williams et al., 1994;Hubbard et al., 1994).
Inhibition of the Ca 2ϩ -ATPase as a Consequence of Anesthetic Binding in Protein Interior-Extensive studies on myoglobin and hemoglobin using NMR spectroscopy measurements of protein in solution and x-ray diffraction studies of crystals saturated with gaseous anesthetics have demonstrated that xenon, cyclopropane, dichloromethane, and halothane bind at discrete sites in the protein interior through van der Waals contacts with nonpolar amino acids and histidines (Schoen-

TABLE II
Apparent quenching constants in the presence of different volatile anesthetics for the purified plasma membrane Ca 2ϩ -ATPase and SR K q was calculated as K q ϭ K app / o ⅐ K app is the slope of the ratio of the total fluorescence intensity versus anesthetic concentration (Fig. 7) as calculated from: F o /F io ϭ 1 ϩ K app [Q], where F o is the total fluorescence intensity in the absence of volatile anesthetic; F io is the total fluorescence intensity at each concentration of anesthetic (both at time 0), and Q is the volatile anesthetic concentration. o is the lifetime of tryptophan in the absence of anesthetic. o for tryptophan in SR was 4.59 ns (Gryczynski et al., 1989) 1969Settle, 1973;Tilton et al., 1984Tilton et al., , 1986Nunes and Schoenborn, 1973). Generally, the binding appears stronger for smaller and more polarizable molecules (Nunes and Schoenborn, 1973). The findings suggest selectivity in the interactions between anesthetics and proteins and a requirement for compatibility between the nonpolar sites available in the protein (proper steric arrangement of amino acid side chains) and the intruding molecule (size, shape, and polarizability). The Ca 2ϩ -ATPase seems to have a fitting nonpolar site(s) available to all five volatile anesthetics studied in our laboratory (present paper and in Kosk-Kosicka and Roszczynska, 1993) and to xenon (Franks et al., 1995) as judged by their inhibitory effects. Temperature effects (on the Ca 2ϩ -ATPase activity) support a nonpolar nature of interactions between the anesthetics and the amino acids in the Ca 2ϩ -ATPase. It would be worthwhile to investigate effects of bigger or/and less polarizable anesthetic molecules on the enzyme to further test our hypothesis.
If volatile anesthetics invade the interior of the Ca 2ϩ -ATPase molecule their binding is expected to affect internal motions and conformational substates of the enzyme (Schoenborn, 1965;Brown et al., 1976;Hibbard and Tulinsky, 1978;Tilton et al., 1984;Lim et al., 1994). The observed effects on total tryptophan fluorescence might then reflect the invasion that results in subtle conformational changes including tryptophan environment. The lack of the Ca 2ϩ -sensitive increase in tryptophan fluorescence in the presence of anesthetics would reflect the increased stability of the substate characteristic of a nonactive enzyme: enzyme with bound anesthetic cannot attain the conformation necessary for its normal activation. As a result no Ca 2ϩ -ATPase activity can be detected. Recent studies, including investigation of the effects of amino acid mutations in the hydrophobic core of -repressor on internal packing interactions, have demonstrated that the protein accommodates the potentially disruptive residues with shifts in its ␣-helical arrangement and becomes more stable since its packing is improved. However, the rearrangements cause repositioning of functional residues, which results in reduced function of the protein (Lim and Sauer, 1991;Lim et al., 1994). By analogy, a distortion of amino acid side chain(s) (resulting from anesthetic binding to the Ca 2ϩ -ATPase) that normally coordinate to Ca 2ϩ may cause a change in Ca 2ϩ binding and this slightly changed substate may not reveal the expected tryptophan fluorescence increase. The small shift observed in Ca 2ϩ affinity at 37°C favors this interpretation as opposed to a total loss of Ca 2ϩ binding in the presence of an anesthetic. Alternatively, insertion of the anesthetic and the resulting rearrangement may affect not Ca 2ϩ binding but the coupling between the Ca 2ϩ binding and induction of a proper conformation that in the active enzyme is detected as the Ca 2ϩ -dependent change in tryptophan fluorescence. Anesthetics effects on the conformational change (normally reflected in Ca 2ϩ -dependent changes in fluorescence of either tryptophan or FITC) correlate well with their effects on the Ca 2ϩ -ATPase activity in support of the interpretation that anesthetics prevent the enzyme from assuming a functionally proper active conformation. This suggestion is consistent with the observation that two activation pathways of the enzyme, by dimerization of enzyme monomers and by calmodulin binding to enzyme monomers, are inhibited by anesthetics to a comparable extent implying a "basic" defect in the activation mechanism. In contrast, the two pathways are affected distinctly different by two other groups of general anesthetics, barbiturates and alkanols (Kosk-Kosicka et al., 1995a). 1 These compounds with the structural properties of detergents are postulated to inhibit Ca 2ϩ -ATPase monomers by binding to an exposed nonpolar patch on the enzyme surface to which calmodulin binds to activate monomers and which is not easily accessible in enzyme dimers. Selectivity of Anesthetic Action on the Ca 2ϩ -ATPase-The reviewed effects on the plasma membrane Ca 2ϩ -ATPase occur at clinical anesthetic concentrations. The Ca 2ϩ -ATPase in sarcoplasmic reticulum is also sensitive to the anesthetics in this concentration range. Its activity is inhibited half-maximally by halothane at an approximately 50% higher concentration than the Ca 2ϩ -ATPase in the plasma membrane ( Fig. 9; see also Kosk-Kosicka and Roszczynska (1993)). Also for this Ca 2ϩ -ATPase the large quenching constants support conformational changes in the presence of anesthetics (Table II). We expect then that the inhibition results from the anesthetic binding in nonpolar sites which correspond to those in the plasma membrane Ca 2ϩ -ATPase (although in SR they may have lower affinity for the anesthetics) and the final consequences of restriction of protein motions are similar in the two Ca 2ϩtransporting enzymes. We have, however, demonstrated a significant difference between the response of the two enzymes The activity assay was performed at 25°C as described under "Materials and Methods." The reaction mixture for sarcoplasmic reticulum contained the divalent cation ionophore A23187 (10 M) and the reaction time was 15 min as compared to 30 min for the plasma membrane Ca 2ϩ -ATPase. For both membranes the Ca 2ϩ -ATPase activity was calculated as the difference between the activity detected in the presence of both Mg 2ϩ and Ca 2ϩ ions and absence of Ca 2ϩ ions. The specific Ca 2ϩ -ATPase activities (in mol of P i /mg of protein/h) were 100 in SR and 0.35 in erythrocyte ghosts.
to lower anesthetic concentrations. The initial activation phase observed only for the SR Ca 2ϩ -ATPase (Fig. 9) which correlates with its effect of the Ca 2ϩ -induced tryptophan fluorescence 4 may be due to changes in its oligomeric size (from bigger to smaller oligomers, as suggested by Karon et al. (1994) based on the time-resolved fluorescence anisotropy data).
In contrast to Ca 2ϩ -ATPase, two other plasma membrane ATPases, the Na ϩ ,K ϩ -ATPase and Mg 2ϩ -ATPase, require at least 3-5 times higher concentrations of volatile anesthetics. 2 Several other membrane proteins that have been reported as potential targets for anesthetics, including voltage-gated sodium, potassium, and calcium channels, are half-maximally inhibited only at halothane concentrations ranging from 4 to 30 times higher than its clinical potency, as summarized in a recent review of the last 10 years of search for the target of anesthetic action (Franks and Lieb, 1994). From among many proteins shown to be affected by the anesthetics only ligandgated ion channels, including GABA receptors and possibly nicotinic acetylcholine receptor, are also targeted at clinical anesthetic concentrations. Studies performed in many laboratories imply GABA A receptors channel complex as a major target for the general anesthetics (for review, see Franks and Lieb (1994)). The pharmacological target is expected to be affected by anesthetics at their clinical concentrations (Ϸ1 MAC) because the dose-response curves for the induction of anesthesia with volatile anesthetics are very steep (concentrations 20% above 1 MAC anesthetize almost all subjects) and because concentrations 2-4 times higher have damaging side effects on the organism. Our findings indicate that the anesthetics action on the Ca 2ϩ -ATPase is selective at their pharmacological concentrations. It remains to be tested whether anesthetic action on the Ca 2ϩ -ATPases and the receptors at the molecular level is exerted by their binding in protein interior. Our hypothesis cannot be verified at present without x-ray structure data.
The finding that volatile anesthetics inhibit the Ca 2ϩ -ATPases in vitro at their clinical concentrations while several other membrane proteins are affected by the anesthetics only at concentrations significantly exceeding those used in the clinical situation opens a possibility that the enzyme is perturbed during anesthesia and this anesthetic action has important physiological consequences. The consequences would be multiple as so many cellular events and cascades depend on precisely controlled intracellular Ca 2ϩ concentrations, ranging from regulation of cell shape in erythrocytes, to neuronal transmission to depression of contractile force in heart.
In conclusion, we propose that the observed correlation between the attenuation of the functionally important conformational change of the enzyme and inhibition of its Ca 2ϩ -ATPase activity could be explained by permeation of the protein molecule by volatile anesthetic and stabilization of its nonactive conformational substate. We view this process as an example of a general phenomena of discriminating binding of small ligands in protein interior.

Maria M. Lopez and Danuta Kosk-Kosicka
The concentrations of volatile anesthetics in the in vitro assays described in our publication are 10-fold higher than indicated. As a result, only the initial activation of the SR Ca 2ϩ -ATPase, in contrast to the inhibition of the plasma membrane Ca 2ϩ -ATPase, occurs at clinical concentrations of volatile anesthetics. The K app calculated from tryptophan quenching experiments using the corrected anesthetic concentrations are about 10-fold lower, i.e. more comparable to the values obtained for other proteins. The concentration error does not affect the main findings of our paper which suggest that the anesthetics interfere with enzymatic function by binding in nonpolar sites in protein interior and modifying conformational substate(s) of the enzyme.

Vol. 271 (1996) 19283-19287
The luteinizing hormone/chorionic gonadotropin receptor has distinct transmembrane conductors for cAMP and inositol phosphate signals.  Fig. 2: We recently discovered that the printed mouse connexin-30 sequence contains an erroneous duplication in the nontranslated part upstream of the coding sequence. The nucleotide sequence from Ϫ1 to Ϫ60 is duplicated in the nucleotide sequence from Ϫ104 to Ϫ163. This is due to an error which occurred during alignment of partial sequences. For correction, the nucleotide sequence Ϫ1 to Ϫ60 has to be deleted so that position Ϫ61 of the printed sequence corresponds to position Ϫ1 of the corrected sequence. The corrected sequence has been deposited in GenBank™ with the accession number Z70023.
We suggest that subscribers photocopy these corrections and insert the photocopies at the appropriate places where the article to be corrected originally appeared. Authors are urged to introduce these corrections into any reprints they distribute. Secondary (abstract) services are urged to carry notice of these corrections as prominently as they carried the original abstracts.