Clotrimazole Inhibits the Ca2+-ATPase (SERCA) by Interfering with Ca2+ Binding and Favoring the E2 Conformation*

Clotrimazole (CLT) is an antimycotic imidazole derivative that is known to inhibit cytochrome P-450, ergosterol biosynthesis and proliferation of cells in culture, and to interfere with cellular Ca2+ homeostasis. We found that CLT inhibits the Ca2+-ATPase of rabbit fast-twitch skeletal muscle (SERCA1), and we characterized in detail the effect of CLT on this calcium transport ATPase. We used biochemical methods for characterization of the ATPase and its partial reactions, and we also performed measurements of charge movements following adsorption of sarcoplasmic reticulum vesicles containing the ATPase onto a gold-supported biomimetic membrane. CLT inhibits Ca2+-ATPase and Ca2+ transport with a KI of 35 μm. Ca2+ binding in the absence of ATP and phosphoenzyme formation by the utilization of ATP in the presence of Ca2+ are also inhibited within the same CLT concentration range. On the other hand, phosphoenzyme formation by utilization of Pi in the absence of Ca2+ is only minimally inhibited. It is concluded that CLT inhibits primarily Ca2+ binding and, consequently, the Ca2+-dependent reactions of the SERCA cycle. It is suggested that CLT resides within the membrane-bound region of the transport ATPase, thereby interfering with binding and the conformational effects of the activating cation.

In addition to biochemical characterization of the ATPase and its partial reactions, we performed direct measurements of charge translocation following adsorption of SR vesicles containing ATPase onto a biomimetic membrane supported by a gold electrode and activation by rapid mixing with a suitable substrate (19). This technique has been used successfully in studies of electrogenic transport by several membrane proteins, such as Na ϩ ,K ϩ -ATPase (20,21), melibiose permease (22), and Na ϩ /proline antiporter (23), including the SERCA pump (24,25). We found that CLT interferes specifically with Ca 2ϩ binding to SERCA even in the absence of ATP. For this reason, the Ca 2ϩ -dependent reactions of this enzyme are also inhibited.
ATPase Preparation-Sarcoplasmic reticulum vesicles were obtained by extraction from the fast-twitch hind leg muscle of New Zealand White rabbit followed by homogenization and differential centrifugation, as described by Eletr and Inesi (27). The vesicles derived from longitudinal SR membrane contained only negligible amounts of the ryanodine receptor Ca 2ϩ channel associated with junctional SR (light vesicles).
Free Ca 2ϩ concentration was calculated with the computer program WinMAXC (28). Unless otherwise stated, 1 M calcium ionophore A23187 was used to prevent Ca 2ϩ accumulation by the SR vesicles (29).
ATPase Activity-Ca 2ϩ -ATPase hydrolytic activity was determined following P i production by a colorimetric method (30 (13). The reaction temperature was 25°C. The reaction mixture was filtrated with an 0.45-m filter, and radioactivity was measured by a liquid scintillation counter.
Phosphorylation by [␥- 32  Phosphorylation by [ 32 P]P i -Enzyme phosphorylation with P i was measured at 25°C in a pH 6.2 reaction mixture containing 50 mM MES, 10 mM MgCl 2 , 2 mM EGTA, and 20% Me 2 SO after a 30-min preincubation with various concentrations of CLT in the presence of 1 mM [ 32 P]P i and 0.5 mg of SR protein/ml. The reaction was acid-quenched after a 10-min incubation by the addition of 1 M perchloric acid. The EP measurements were conducted using a filtration method.
Measurement of Charge Movements-Charge movements were measured by adsorbing the SR vesicles containing the Ca 2ϩ -ATPase onto a biomimetic membrane supported by a gold electrode (the socalled solid-supported membrane (SSM)). The SSM consisted of an octadecanethiol monolayer covalently bound to a gold surface via the sulfur atom with a phospholipid monolayer on top of it (31). After adsorption, usually carried out at an applied potential of ϩ0.1 V, the protein was activated by a rapid injection of a solution containing the appropriate substrate, e.g. ATP or Ca 2ϩ ions. If at least one electrogenic step is involved in the relaxation process, a current transient (pre-steady state) can be recorded along the external circuit. The acquisition of transients obtained under different experimental conditions, together with their subsequent elaboration, can provide important kinetic information about protein function and/or its modulation by drugs, peptides, and small soluble proteins. In particular, from a current transient, three main data can be extracted: the peak current, the translocated charge, and the decay time constants. The charge is obtained by numerical integration of the current transient, whereas the decay time constants are provided by a multiexponential fitting of the transient (19,20,24,25). The time constants are strictly correlated with the rate constants of the reaction steps involved in the activation of the pump (32). The preparation of the gold electrodes and the whole experimental set-up as well as the solution exchange technique are described in detail by Tadini-Buoninsegni et al. (25).
For the activation experiments, the composition of the solutions was the following:

RESULTS
We studied the effect of CLT on the SR Ca 2ϩ -ATPase by a series of biochemical measurements, including ATPase activity, [ 45 Ca]Ca 2ϩ binding in the absence of ATP, formation of the phosphorylated enzyme intermediate (E-P) by utilization of [␥-32 P]ATP in the presence of Ca 2ϩ , and formation of E-P by utilization of [ 32 P]P i in the absence of Ca 2ϩ . In addition, we performed pre-steady state activation experiments with the SSM-based technique (Ca 2ϩ concentration jumps and ATP concentration jumps in the presence of CLT).
Biochemical Experiments-It is shown in Fig. 1 that the steady state hydrolytic activity of the Ca 2ϩ -ATPase (in the presence of 10 or 200 M free Ca 2ϩ ) is inhibited by CLT within the M concentration range. It is noteworthy that measurements at two different Ca 2ϩ concentrations (10 and 200 M) yield the same dependence upon the inhibitor concentration, indicating that no protection is obtained by increasing the Ca 2ϩ concentration above the saturating level.
With the aim of determining whether the effect of CLT is a total inhibition of SERCA activity (as that of TG) or is rather because of interference with a specific partial reaction, we measured independently [ 45 Ca]Ca 2ϩ binding in the absence of ATP, E-P formation by utilization of [␥-32 P]ATP in the presence of Ca 2ϩ , and E-P formation by utilization of [ 32 P]P i in the absence of Ca 2ϩ . It is shown in Fig. 2 that 45 Ca 2ϩ binding and E-P formation from [␥-32 P]ATP are inhibited with a similar pattern as the concentration of CLT is increased from zero to 50 M. It should be pointed out that a limit in the solubility of CLT in the aqueous reaction medium renders a concentration increase above 50 M rather ineffective. It is noteworthy that the same pattern of inhibition with respect to the CLT concentration was observed when enzyme phosphorylation with ATP was measured in the presence of 10, 50, or 200 M Ca 2ϩ , indicating that the CLT inhibition is not competitive with respect to Ca 2ϩ . A most interesting observation is that E-P formation from [ 32 P]P i (in the absence of Ca 2ϩ ) is only minimally reduced (Fig. 2), indicating that CLT inhibition of the SERCA pump is mainly exerted on Ca 2ϩ binding and the Ca 2ϩ -dependent partial reactions of the enzymatic cycle.

On and Off Ca 2ϩ Concentration Jumps in the Presence of CLT in
Pre-steady State-A Ca 2ϩ concentration jump, performed with the SSM-based technique, yielded a current transient that could be ascribed to Ca 2ϩ binding to the ATPase. An ON signal was obtained following the injection of the Ca 2ϩ -containing solution (Fig. 3A, step 1), and an OFF signal was obtained upon calcium release from the binding site following injection of a solution containing EGTA and no calcium (Fig.  3A, step 2) (25). We then performed these experiments in the presence of increasing concentrations of CLT to assess the effect of the drug on Ca 2ϩ binding in the absence of ATP. In agreement with direct measurements of Ca 2ϩ binding (Fig. 2), we found that charge movements attributed to Ca 2ϩ binding were also reduced by CLT. Fig. 4 compares a saturating calcium concentration jump in the absence and in the presence of 5 M CLT, showing a clear reduction of the signal in the presence of the drug. Moreover, a titration with increasing Ca 2ϩ concentrations (Fig. 5) shows that the signal is reduced to about one-third of that recorded in the absence of the inhibitor (25). On the other hand, despite the reduction in maximal binding, the K d and n values (0.53 and 1.7 M, respectively) obtained from fitting to the Hill function are very close to those obtained in the absence of drug, either by measurements of charge movements (25) or by direct measurements with a radioactive tracer (33). This indicates that non-saturating CLT reduces the amount of Ca 2ϩ bound but does not affect the binding characteristics (i.e. affinity) of the enzyme molecules that are not totally inhibited.
We also found that titrations with increasing CLT concentrations in the presence of 10 or 100 M Ca 2ϩ concentrations (Fig. 6) yield the same inhibition curve. These results are in agreement with the experiments on enzyme phosphorylation with ATP (Fig. 2) and ATPase hydrolytic activity (Fig. 1) and imply that increasing the Ca 2ϩ concentration above the saturating level does not protect the enzyme from CLT. Therefore, the CLT inhibition is not competitive with respect to Ca 2ϩ . On the other hand, we found that the effective CLT concentration is lower when the inhibitor is added before (K I ϭ 7 M, as in Fig. 6) rather than after Ca 2ϩ (K I ϭ 35 M, as in Figs. 1, 2,  and 7). A similarly greater resistance of the Ca 2ϩ -bound conformation of the enzyme (Ca 2 E1) as compared with the Ca 2ϩ -free conformation (E2) was found previously with respect to TG (13) and miconazole (34).
ATP Concentration Jumps in the Presence of CLT-An ATP concentration jump in the presence of Ca 2ϩ , performed on a SSM with SR vesicles containing Ca 2ϩ -ATPase adsorbed onto it, produces a current transient because of the electrogenicity of Ca 2ϩ release into the lumen of the SR and, to some extent, to Ca 2ϩ re-binding to the protein after the hydrolytic cleavage of P-E2 intermediate (Fig. 3B, steps 2 and 4). It was previously determined that the formation and hydrolytic cleavage of the phosphorylated intermediate are not electrogenic (35). Fig. 7 shows the effect of 35 M CLT on a 100 M ATP concentration jump in the presence of saturating Ca 2ϩ (100 M). It is clear that the current transient because of Ca 2ϩ translocation is affected by CLT. In fact, starting from no CLT and increasing CLT concentration up to 35 M, peak currents are progressively decreased. Furthermore, integration of the current transients yields values for moved charges that exhibit a behavior very similar to that of peak currents with respect to the effect of CLT (Fig. 7, inset).
It is noteworthy that exponential fitting of the current transients yields two time constants for current decay. The two constants, 1 and 2 , have average values of 27 Ϯ 2 and 42 Ϯ 3 ms, respectively and do not appear to be significantly affected by CLT (not shown), irrespective of the effect of CLT on peak currents and moved charge.

DISCUSSION
CLT is an antimycotic agent, the effect of which is attributed to inhibition of cytochrome P-450 and ergosterol biosynthesis in yeast (1,2). It is also shown that CLT inhibits cell proliferation in vitro (3), interferes with calcium homeostasis in culture cells (3,4), and inhibits the calcium pump (SERCA2) obtained with the microsomal fraction of heart muscle (5). We have characterized in detail the effect of CLT on the Ca 2ϩ -ATPase (SERCA1) activity associated with microsomal vesicles obtained from skeletal muscle. This enzyme preparation is highly purified and well suited to characterization of the catalytic cycle and its partial reactions (6 -9). In addition, several SERCA inhibitors have been described (13)(14)(15)(16)(17)(18), thereby yielding useful grounds for comparison of inhibitory mechanisms. Finally, the simultaneous use of biochemical and biophysical methods in the characterization of the CLT inhibitory mechanism has given us the opportunity to demonstrate clearly the functional relevance of charge transfer measurements and their correspondence to partial reactions of the ATPase cycle.
As outlined in Fig. 3B, the ATPase (SERCA) cycle requires initial activation through Ca 2ϩ binding from the exterior of the vesicles (step 4) followed by utilization of ATP to form a phosphorylated intermediate (step 1). The bound Ca 2ϩ is then dissociated into the lumen of the vesicles (Fig. 3B, step 2) followed by hydrolytic cleavage of the phosphoenzyme (step 3). Using direct binding assays by a radioactive tracer as well as measurements of charge movements by the SSM-based technique, we found that the initial Ca 2ϩ binding, before utilization of ATP, is inhibited by CLT (Figs. 2, 4, and 6). It is of interest that plots of charge as a function of Ca 2ϩ concentration (Fig. 5) yield affinity constants and cooperative behavior identical to that of equilibrium binding isotherms obtained in the absence of CLT by direct measurements with isotopic tracer (33). The correspondence of the two types of measurements is highly satisfactory and demonstrates the specificity of charge movement with regard to Ca 2ϩ binding and its inhibition by CLT.
Considering the absolute Ca 2ϩ requirement for enzyme activation, it is expected that steady state ATPase activity will be inhibited in parallel with inhibition of Ca 2ϩ binding by CLT. In fact, we found that steady state ATPase activity and Ca 2ϩ transport were inhibited by the same CLT concentrations producing inhibition of Ca 2ϩ binding. This was demonstrated by measuring P i by a colorimetric method after adding ATP to the enzyme in the presence of Ca 2ϩ (Fig. 1) as well as by measuring charge translocation following an ATP concentration jump by the SSM-based technique (Fig. 7). Here again, identical results were obtained by the biochemical and biophysical methods. An interesting observation is that, whereas CLT clearly decreases the amount of moved charge (Fig. 7), the decay kinetics of the remaining current transient is not affected by CLT (not shown). Therefore, although an increasing number of enzyme molecules are inactivated by CLT, the remaining enzyme molecules do not bind CLT and continue to function normally. This suggests a specific inhibitory site for CLT within the ATPase molecule whereby the ATPase molecules are inhibited independent of each other. Furthermore, comparative experiments performed by addition of CLT before or after Ca 2ϩ suggest that the Ca 2ϩ -free conformation of the enzyme is more sensitive to CLT and is stabilized by the bound inhibitor.
We also found that phosphoenzyme formation by utilization of ATP in the presence of Ca 2ϩ is inhibited by CLT within the same concentration range as Ca 2ϩ binding (Fig. 2) and ATPase hydrolytic activity (Fig.   1), with a pattern consistent with noncompetitive inhibition. However, phosphoenzyme formation by utilization of P i in the reverse direction of cycle is only minimally affected by CLT (Fig. 2), as also observed by Snajdrova et al. (5). This reverse reaction is not dependent on activation by Ca 2ϩ and in fact requires removal of bound Ca 2ϩ (36). Therefore, it is apparent that the ATPase catalytic mechanism is not primarily affected by CLT but only secondarily, through interference with activation by Ca 2ϩ binding. Such a specific interference with a partial reaction (i.e. Ca 2ϩ binding) is of interest when compared with the global inhibition of the ATPase cycle produced by TG. So far, 1,3-dibromo-2,4,6-tris(methylisothiouronium)benzene is the only other ATPase inhibitor producing specific inhibition of a single partial reaction, i.e. the E1-P 7 E2-P transition (37).
It should be noted that CLT shows a rather high affinity for SERCA with a K I of 35 M in the presence of Ca 2ϩ and K I of 7 M in the absence of Ca 2ϩ . Considering the hydrophobic character of the CLT molecule and the location of cation-binding sites in the membrane-bound region of transport ATPases, it is apparent that CLT interferes at this location with the binding of the activating cation and with the conformational change involved in cation binding. In addition to the significance of the CLT effect with regard to the catalytic and possibly pharmacological mechanism, we consider that stabilization of the Ca 2ϩ -free, E2 conformation of the enzyme, leaving the catalytic site in a functional state as revealed by the P i reaction, may be an extremely useful tool in structural and crystallization studies of cation transport ATPases. In fact, ATPase crystal structures of the highest resolution are presently obtained by incorporation of two (rather than one) inhibitors, permitting detection of ordered water molecules and advanced electrostatic calculations yielding estimates of the proton occupancy of acidic residues (38). At the same time, the availability of several inhibitors supports organic synthesis of new, highly specific and potent compounds, guided by the available crystal structure of the ATPase (39,40).