This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Buoninsegni, F. T. Articles by Guidelli, R. Search for Related Content PubMed PubMed Citation Articles by Buoninsegni, F. T. Articles by Guidelli, R.

Time-Resolved Charge Translocation by Sarcoplasmic Reticulum Ca-ATPase Measured on a Solid Supported Membrane

Francesco Tadini Buoninsegni ^*, Gianluca Bartolommei ^*, Maria Rosa Moncelli ^*, Giuseppe Inesi  and Rolando Guidelli ^*

^* Department of Chemistry, University of Florence, 50019 Sesto Fiorentino, Italy; and Department of Biochemistry and Molecular Biology, University of Maryland School of Medicine, Baltimore, Maryland 21201 USA

Correspondence: Address reprint requests to Prof. Rolando Guidelli, Dept. of Chemistry, University of Florence, via della Lastruccia 3, 50019 Sesto Fiorentino, Italy. Tel.: 39-055-4573097; Fax: 39-055-4573098; E-mail: guidelli{at}unifi.it.

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSIONS APPENDIX ACKNOWLEDGEMENTS REFERENCES

 
Sarcoplasmic reticulum vesicles were adsorbed on an octadecanethiol/phosphatidylcholine mixed bilayer anchored to a gold electrode, and the Ca-ATPase contained in the vesicles was activated by ATP concentration jumps both in the absence and in the presence of K⁺ ions and at different pH values. Ca²⁺ concentration jumps in the absence of ATP were also carried out. The resulting capacitive current transients were analyzed together with the charge under the transients. The relaxation time constants of the current transients were interpreted on the basis of an equivalent circuit. The current transient after ATP concentration jumps and the charge after Ca²⁺ concentration jumps in the absence of ATP exhibit almost the same dependence upon the Ca²⁺ concentration, with a half-saturating value of 1.5 µM. The pH dependence of the charge after Ca²⁺ translocation demonstrates the occurrence of one H⁺ per one Ca²⁺ countertransport at pH 7 by direct charge-transfer measurements. The presence of K⁺ decreases the magnitude of the current transients without altering their shape; this decrease is explained by K⁺ binding to the cytoplasmic side of the pump in the E₁ conformation and being released to the same side during the E₁–E₂ transition.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSIONS APPENDIX ACKNOWLEDGEMENTS REFERENCES

 
The Ca-ATPase of the sarcoplasmic reticulum (SR) is an integral membrane protein that couples the hydrolysis of a molecule of ATP to the active transport of two Ca²⁺ ions across the membrane of SR (Møller et al., 1996; Lee and East, 2001). It plays an essential role in regulating intracellular calcium concentration, which is kept at or below 0.1 µM by pumping ions from the cytoplasm into the SR lumen: in this manner, SR Ca-ATPase induces muscle relaxation, contributing to calcium homeostasis. Reduced activity of this pump may result in prolonged elevated calcium levels, which may lead to stiffness and muscle relaxation problems (Gommans et al., 2002; MacLennan, 2000).

Like other members of the P-type ATPases class, Ca-ATPase forms an aspartyl-phosphate intermediate during the enzymatic reaction cycle. According to the E₁–E₂ model, Ca²⁺ binding to the cytosolic domain in the high-affinity E₁ conformation is followed by the phosphorylation of an aspartyl residue (Asp-351) by ATP. A conformational change of the phosphoenzyme from the E₁P to the E₂P state exposes Ca²⁺ to the lumenal side and promotes Ca²⁺ release, due to the low Ca²⁺ affinity for the E₂P state of the pump. After the cleavage of the phosphoenzyme, the pump returns to the E₁ conformational state (Carafoli and Brini, 2000). The recent availability of the crystal structure of SR Ca-ATPase, both in the E₁Ca₂ conformation (Toyoshima et al., 2000) and in the E₂ conformation stabilized by thapsigargin (Toyoshima and Nomura, 2002), has represented a breakthrough in the understanding of structure-function relationships (Lee and East, 2001; Hua et al., 2002).

From functional studies, Ca²⁺ binding to Ca-ATPase was found to be electrogenic (Butscher et al., 1999). Kinetic studies have shown that under physiological conditions Ca-ATPase couples the translocation of two Ca²⁺ ions to the countertransport of two protons (Yu et al., 1993, 1994). However, the Ca²⁺/H⁺ stoichiometry changes when the cytoplasmic pH is varied at constant lumenal pH or when the lumenal pH is varied at constant cytoplasmic pH (Yu et al., 1994).

Potassium ion was reported to influence the pumping activity of Ca-ATPase in different ways. The construction of chimeric ATPases between Na,K-ATPase and Ca-ATPase revealed two distinct regions with different K⁺ affinity (Ishii et al., 1997; Yoshimura et al., 1997). A potassium binding site that induces a decrease in Ca²⁺ affinity for the pump, when occupied by K⁺ ions, was proposed by Lee et al. (1995) on the basis of kinetic studies. A competition of K⁺ with Ca²⁺ for the Ca²⁺ binding site less exposed to the cytoplasm was also postulated (Lee et al., 1995). Potassium ion was also reported to accelerate the decomposition of the ADP-insensitive form of the enzyme (Yamada and Ikemoto, 1980) and the slow component of the biphasic fluorescence signal due to Ca²⁺ binding, when in the presence of Mg²⁺ (Moutin and Dupont, 1991).

Presteady-state electrical measurements of the activity of an ion pump yield direct information about the time dependence of the charge movement across the pump (Läuger, 1991; Pintschovius and Fendler, 1999). Adsorbing membrane fragments or proteoliposomes on a conventional black lipid membrane (BLM) (Hartung et al., 1987, 1997) or on a derivatized solid support (Pintschovius and Fendler, 1999) and activating them by a concentration jump causes a certain distortion of the pump current, due to the combined support-membrane system; this, however, can be satisfactorily accounted for. A convenient method to perform concentration jumps of an arbitrary substrate at the surface of a solid-supported membrane (SSM) was devised by Pintschovius and Fendler (1999). The SSM consists of an alkanethiol monolayer firmly anchored to a gold surface via the sulfhydryl group, with a second phospholipid monolayer on top of it. Membrane fragments or proteoliposomes are then adsorbed on this gold-supported mixed thiol-lipid bilayer. This technique combines the high mechanical stability of the SSM with a rapid solution exchange procedure. This method has been successfully used to investigate the electrogenic partial reactions in the enzymatic cycle of Na,K-ATPase (Pintschovius and Fendler, 1999; Pintschovius et al., 1999; Tadini Buoninsegni et al., 2000; Tadini Buoninsegni et al., 2003) and the charge transfer of the melibiose permease (Ganea et al., 2001).

This study describes an application of the SSM technique to the investigation of the pumping activity of SR Ca-ATPase. SR vesicles containing ATPase from rabbit skeletal muscle were adsorbed on the SSM. Upon adsorption, the ion pumps were activated by ATP concentration jumps at variable ATP concentration, Ca²⁺ concentration, and pH, and the current transients generated by Ca-ATPase activity were measured under potentiostatic conditions. Ca²⁺ concentration jumps in the absence of ATP were also carried out, to investigate Ca²⁺ binding to and release from the pump. Finally, the influence of the presence of physiological concentrations of K⁺ ions on the Ca-ATPase pumping activity was studied.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSIONS APPENDIX ACKNOWLEDGEMENTS REFERENCES

 
Chemicals
Calcium, potassium and magnesium chlorides, and Tris were obtained from Merck (Darmstadt, Germany) at analytical grade. Adenosine-5'-triphosphate disodium salt (ATP, 97%) and dithiothreitol (DTT, 99%) were purchased from Fluka (Buchs, Switzerland). Octadecanethiol (98%) from Sigma-Aldrich (St. Louis, MO) was used without further purification. EGTA, tetraethylammonium chloride (TEA, 98%), thapsigargin, and calcimycin (calcium ionophore A23187) were obtained from Sigma-Aldrich. Protonophore 1799, (2,6-dihydroxy)-1,1,1,7,7,7-hexafluoro-2,6-bis(trifluoro-methyl)heptane-4-one, was kindly provided by the Max-Planck-Institüt für Biophysik.

The lipid solution contained diphytanoylphosphatidylcholine (Avanti Polar Lipids, Alabaster, AL) and octadecylamine (puriss., Fluka) (60:1) and was prepared at a concentration of 1.5% (w/v) in n-decane (Merck) as described by Bamberg et al. (1979).

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 (1972). The vesicles so obtained, derived from longitudinal SR membrane, contained only negligible amounts of the ryanodine receptor Ca²⁺ channel associated with junctional SR (light vesicles). The protein/lipid ratio was 1:1 and the total protein content was 22.4 mg/ml, of which 50% consisting of Ca-ATPase.

The free calcium concentration was calculated with the computer program Winmaxc v. 2.40 (Bers et al., 1994). Unless otherwise stated, 1 µM calcium ionophore A23187 was used to prevent Ca²⁺ accumulation inside the vesicles (Hartung et al., 1997).

The solid supported membrane
The SSM consisted of an alkanethiol monolayer covalently bound to a gold surface via the sulfur atom, with a phospholipid monolayer on top of it (Seifert et al., 1993; Florin and Gaub, 1993). To prepare the SSM, the procedure described by Pintschovius and Fendler (1999) was followed. Briefly, the mixed bilayer was formed in two sequential self-assembly steps. A self-assembled octadecanethiol monolayer was first formed on a gold electrode by incubating a freshly deposited gold film in an ethanol solution of 1 mM octadecanethiol for 6 h at room temperature. The bilayer was then formed by spreading a drop of lipid solution (usually 5 µl) on the surface of the thiol-coated gold electrode. Typically, the effective membrane area ranged from 2 to 3 mm².

Setup
To carry out rapid concentration jumps, a Plexiglas cuvette with an inner volume of 20 µl was used. The SSM and an O-ring, which contained the actual solution volume, were sandwiched between the upper and the lower part of the cuvette. The SSM acted as the working electrode, while an Ag/AgCl(0.1M KCl) electrode was employed as a counterelectrode. The counterelectrode was separated from the streaming solution by an agar/agar gel bridge. For details, see Pintschovius and Fendler (1999).

Two different 100 ml glass containers were used for the nonactivating and the activating solution. Unless otherwise stated, the activating solution differed from the nonactivating one only by the presence of the species activating the pump or binding to it. When performing a concentration-jump experiment, the solution flow was kept constant at 60 ml/min by applying a pressure of 0.4 bar to the system and by controlling the pressure with a precision digital manometer. The cuvette was connected to the outlet of a Teflon block on which two solenoid valves were mounted (model 225T052, NResearch, West Caldwell, NJ). The two valves, which were computer controlled through a digital-to-analog converter (DAC 488/2, IOtech, Cleveland, OH), allowed a fast switching between the activating and the nonactivating solution. All parts of the setup conducting the electrolyte solutions were enclosed in a Faraday cage. The current, generated by the ion pumps upon keeping the applied potential between the SSM and the counterelectrode equal to zero, was amplified by a current amplifier (Keithley (Cleveland, OH) 428, gain: 10⁹ V/A), filtered (low-pass, 3 ms), recorded (16-bit analog-to-digital converter, IOtech ADC 488/8SA), visualized (Oscilloscope, Tektronix (Beaverton, OR) TDS 340A) and stored (Power PC G3, Macintosh, Apple, Cupertino, CA). Operation of the experimental setup and data acquisition were carried out under computer control (GPIB interface, National Instruments (Austin, TX) board) using a homemade acquisition program written in LabView (National Instruments) environment.

Solution exchange technique
Two hours after forming the SSM and filling the cuvette, the capacitance and conductance of the SSM remained constant at C_m = 0.2–0.4 µF/cm² and G_m = 50–100 nS/cm². At this stage of the procedure, control experiments were usually performed with the protein-free SSM to exclude any artifacts generated by the solution exchange (Pintschovius and Fendler, 1999). The SR vesicles containing Ca-ATPase were then added by injecting 20 µl of their suspension into the cuvette through the outlet opening. The suspension was then vigorously mixed using a pipette. The vesicles were adsorbed on the SSM for 30 min upon applying a potential difference of +0.1 V. The usual procedure for a concentration-jump experiment consisted of three steps: i), washing the cuvette with the nonactivating solution for 1 s; ii), injecting the activating solution into the cuvette for 1 s; and iii), removing the activating solution from the cuvette with the nonactivating solution for 1 s.

To verify the reproducibility of the current transients generated within the same set of measurements on the same SSM, each single measurement of the set was repeated 4–5 times and then averaged to improve the signal/noise ratio. Average standard deviations were usually found to be no >±5%. At the beginning of each set of measurements, 100 µM ATP jumps were carried out to test the activity of the ion pump previously adsorbed on the gold-supported alkanethiol/phospholipid mixed bilayer. The same ATP jump was performed at the end of the set of measurements, and the initial and final ATP-induced current transients were then compared to rule out any loss of activity during the time of the experiment. If differences between the two transients were >±5%, the set was discarded.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSIONS APPENDIX ACKNOWLEDGEMENTS REFERENCES

 
Electrical currents generated by Ca-ATPase were measured by adsorbing native vesicles containing Ca-ATPase from rabbit skeletal muscle on the SSM. The calcium pumps were then activated under different experimental conditions.

Different ATP concentration jumps at a constant, saturating Ca²⁺ concentration
Fig. 1 shows a typical potentiostatic current transient after a 100 µM ATP concentration jump in the presence of a free calcium concentration of 100 µM. The sign of the current peak is negative and corresponds to the transport of positive charge from the aqueous solution toward the SSM (Dolfi et al., 2002). The direction of the current indicates that the native vesicles containing Ca-ATPase that contribute to the electrical signal are adsorbed with the cytoplasmic side facing the aqueous solution.

View larger version (17K): [in this window] [in a new window]   FIGURE 1  Current transient after an ATP concentration jump obtained with a nonactivating solution containing 150 mM choline chloride, 1 mM MgCl2, 1.1 mM CaCl2 (100 µM free calcium), 25 mM TRIS, 1 mM EGTA, and 0.2 mM DTT at pH 7.0 (HCl). The activating solution had the same composition as the nonactivating one plus ATP at a saturating concentration of 100 µM. Solid curve is the best fit of the experimental curve to the biexponential function A1exp(–t/1) + A2exp(–t/2) upon setting t = 0 at tpeak. Inset shows the plot of Ipeak (normalized, see text) versus. cpeak for various ATP concentrations; the solid curve is the best fit to the Michaelis-Menten equation.

(1)

where t_peak is the time of the current peak, as measured from the onset of the current rise, and _app is an empirical parameter, which can be determined as described in Pintschovius and Fendler (1999), provided the ATP dependence of the peak current, I_peak, satisfies the Michaelis-Menten equation. (Note that the above authors measure the time from the instant of the electrical signal that activates the electrical valve.) In this case, the best fit of the experimental data was obtained for _app = 92 ms and for a half-saturating concentration K_M = 2.9 µM. The confidence interval for _app was between 85 and 97 ms. The inset of Fig. 1 shows the experimental plot of I_peak versus c_peak for various ATP concentrations, c₀; the solid curve is the best fit to the Michaelis-Menten equation. The experimental points were obtained from two sets of current transients recorded on two different SSMs. The error bars express the average standard deviations in the 4–5 repeated measurements routinely carried out on the same SSM. Since the amount of adsorbed vesicles varies from a SSM to another, the peak currents of each set were normalized to their maximum value recorded under ATP saturating conditions, taken as unity. For a detailed analysis, the descending portion of the current transients was fitted with the biexponential function A₁exp(–t/₁) + A₂exp(–t/₂), upon setting t = 0 at t_peak (solid curve in Fig. 1). Fig. 2 shows plots of and A₁ versus c_peak, as obtained from a single set of current transients. It is apparent that A₁, and I_peak exhibit approximately the same dependence upon the ATP concentration. The second relaxation time constant, ₂, is practically independent of c_peak (data not shown), and amounts to 300 ms, whereas its amplitude A₂ is positive and more than one order of magnitude smaller than the maximum absolute value of A₁. The second exponential function accounts for the current overshoot, which is evident in the current transient of Fig. 1. The charge under any of the current transients recorded on the same SSM is practically the same for all ATP concentrations, and corresponds to the overall amount of Ca²⁺ ions translocated by the pumps in a cycle. No stationary current is observed, due to the high resistance of the supporting alkanethiol/phospholipid mixed bilayer (see the Appendix).

View larger version (10K): [in this window] [in a new window]   FIGURE 2  Plot of the first relaxation time constant, 1 (), and of the corresponding amplitude, A1 (), versus the corrected ATP concentrations, cpeak, under the same conditions as in Fig. 1.

100 µM ATP concentration jumps at different Ca²⁺ concentrations
If jumps of a saturating ATP concentration of 100 µM are carried out in the presence of various Ca²⁺ concentrations both in the nonactivating and in the activating solution, the resulting peak currents depend upon the Ca²⁺ concentration as shown in the semilogarithmic plot of Fig. 3. No correction of the concentration values was required in this mode of concentration-dependent measurement. In fact, even if the ATP concentration in contact with the SSM at the current peak is lower than its full value in view of Eq. 1, it is still high enough to assure saturation of the calcium pumps with the cytoplasmic side facing the aqueous solution.

View larger version (8K): [in this window] [in a new window]   FIGURE 3  Ca2+ dependence of 100 µM ATP concentration-jump experiments. The solution contained 150 mM choline chloride, 1 mM MgCl2, 25 mM TRIS, 1 mM EGTA, and 0.2 mM DTT at pH = 7.0 (HCl), and free calcium concentrations ranging from 0.6 to 53.7 µM; the latter were realized by suitable additions of CaCl2. Solid curve is the best fit to Eq. 2.

(2)

Here, Z is the ratio of the free calcium concentration to its experimental half-saturating value, K_1/2 = 1.26 ± 0.08 µM, whereas K₁ and K₂ are the binding constants for the first and second Ca²⁺ ion. The parameter R measures any cooperativity between the two ions. When R equals unity, the two ions bind independently from each other, and the binding isotherm reduces to a Langmuir isotherm. In this case, K₁ equals 4K₂, because there are four possibilities for any of the two binding sites being occupied by any of the two still unbound Ca²⁺ ions. Once one of the two binding sites is occupied by one of the two Ca²⁺ ions, the remaining Ca²⁺ ion is left with only one possibility. The best fit to the experimental points in Fig. 3 was obtained for R = 1.3 ± 0.3. The R value being greater than unity denotes a cooperative binding, in qualitative agreement with the literature. The two binding constants K₁ and K₂, derived from Eq. 2 for K_1/2 = 1.26 µM and R = 1.3, amount to 1.2 x 10⁶ M^–1 and 5.1 x 10⁵ M^–1, respectively.

Ca²⁺ concentration jumps in the absence of ATP
In Fig. 4, the current transient after a saturating Ca²⁺ concentration jump ([Ca²⁺]_free = 28.2 µM) in the absence of ATP (curve a) is compared with that after a 100 µM ATP concentration jump in the presence of the same Ca²⁺ concentration, under otherwise identical conditions (curve b). The charge under the first current transient, due to Ca²⁺ binding to the pump, is less than the charge under the second transient, due to the ATP-induced Ca²⁺ translocation, by a factor of 0.63. Here and in the following, the charge under the current transients was calculated upon eliminating the contribution from the current overshoot. To this end, the whole charge enclosed between the current transient and the time axis was first measured, including its positive contribution due to the current overshoot. The time axis was taken as the horizontal axis passing by the background current attained toward the end of the 1 s period of exposition of the SSM to the activating solution. This overall negative charge was then increased by the negative quantity A₂₂, where ₂ is the relaxation time for the overshoot and A₂ is the corresponding positive amplitude. (See the Discussion section for a justification of this procedure.) For the small A₂₂ values normally observed, practically identical results were obtained by measuring only the charge on the negative side of the current axis. Henceforth, the current after the rapid injection of an activating solution into the cuvette will be referred to as the on-current, whereas the current after the subsequent rapid injection of a nonactivating solution will be referred to as the off-current. Clearly, the currents so far described are on-currents.

View larger version (17K): [in this window] [in a new window]   FIGURE 4  Current transients after a 28.2 µM free Ca2+ concentration jump in the absence of ATP (a) and a 100 µM ATP concentration jump in the presence of 28.2 µM free Ca2+ in both the activating and the nonactivating solution (b). The nonactivating solution contained 150 mM choline chloride, 1 mM MgCl2, 25 mM TRIS, 0.2 mM EGTA, and 0.2 mM DTT at pH 7.0 (HCl); in the case of curve b, it also contained 28.2 µM free Ca2+. Inset shows the on-current after the Ca2+ concentration jump in the absence of ATP, as well as the subsequent off-current after the rapid displacement of the Ca2+-containing solution by the nonactivating solution.

As expected, the charge under the on-current transient was found to be equal and opposite to that under the corresponding off-current transient at all Ca²⁺ concentrations. In fact, whereas the first transient is due to Ca²⁺ binding to the pump, the second one is due to its removal. Fig. 5 shows a plot of the charge under the off-current transient versus the free calcium concentration present in the preceding activating solution. The fit of the experimental points with the Hill function yields a half-saturating Ca²⁺ concentration of 1.5 ± 0.3 µM and a Hill coefficient of 1.1 ± 0.2.

View larger version (9K): [in this window] [in a new window]   FIGURE 5  Plot of the charge under the off-current transients after Ca2+ concentration jumps against the free calcium concentration. The nonactivating solution was the same as in Fig. 4. The free calcium concentration in the activating solution was varied from 0.317 to 42.2 µM by varying the total calcium concentration from 82.5 µM to 0.24 mM. Solid curve is the best fit of the experimental points to the phenomenological Hill function.

View larger version (12K): [in this window] [in a new window]   FIGURE 6  Dependence of off,1 () and off,2 () on the free calcium concentration. off,1 and off,2 are the relaxation time constants of the off-current transients obtained under the same experimental conditions as in Fig. 5.

pH dependence of charge translocation at a constant, saturating Ca²⁺ concentration
Fig. 7 shows a series of current transients after 100 µM ATP concentration jumps in the presence of a free calcium concentration of 100 µM and at different pH values. In this experiment both the calcium ionophore A23187 and the protonophore 1799 were used, to prevent the formation of Ca²⁺ and H⁺ gradients across the membrane and to reduce the transmembrane potential. The presence of these two ionophores determined the attainment of a stable stationary "pump" current, which was revealed by an appreciable capacitive off-current flowing from the electrode toward the solution, namely in the opposite direction with respect to the on-current. It should be noted that a nonzero stationary pump current does not necessarily involve the flow of a nonzero stationary "on-current" along the external circuit. In fact, due to the high resistance of the supporting alkanethiol/phospholipid mixed bilayer, the current transients in Fig. 7 do not show a detectable stationary on-current.

View larger version (21K): [in this window] [in a new window]   FIGURE 7  Current transients after 100 µM ATP concentration jumps at different pH values: 6.55 (*), 6.78 (+), 7.03 (), 7.35 (), 7.58 (), and 8.13 (•). The nonactivating and activating solutions had the same composition as in Fig. 1. Both the calcium ionophore A23187 and the protonophore 1799 (1.25 µM) were used. Inset shows the dependence of the normalized charge QN under the peaks upon pH. Solid curve is the best fit of the experimental points to the phenomenological Hill function.

View larger version (11K): [in this window] [in a new window]   FIGURE 8  Plots of the peak current, Ioff (), and of the single relaxation time constant, off (), of the off-current transients against pH. Data were obtained from the set of current transients in Fig. 7.

(3)

yielding a K_decay value of 18 ± 4 mM. For [K⁺] 100 mM, the charge under the current transient decreases with increasing [K⁺], in the same way as the peak current does; moreover, the decreasing branch of the current transient can be satisfactorily fitted by a single exponential function, with a relaxation time constant of 15 ms. These results agree with those of Hartung et al. (1997), who observed that the presence of 50 mM K⁺ causes the amplitude of the current transient after an ATP concentration jump on SR fragments adsorbed on a BLM to be reduced by 50%, whereas the time course is nearly unchanged. From Fig. 9 it is apparent that for [K⁺] > 100 mM, the current transients show a tail than lasts for almost 40 ms.

View larger version (22K): [in this window] [in a new window]   FIGURE 9  Current transients after 100 µM ATP concentration jumps in the presence of different K+ concentrations: 0 (•), 10 (), 20 (), 100 (), 200 (+), and 225 (*) mM. Currents were normalized to the maximum peak current, taken as unity. The nonactivating solution contained x mM KCl, with 0 < x < 225 mM, (250 – x) mM choline chloride, 1 mM MgCl2, 0.225 mM CaCl2 (28 µM free calcium concentration), 25 mM TRIS, 0.2 mM EGTA, and 0.2 mM DTT at pH 7.0 (HCl). The activating solution had the same composition as the nonactivating one plus 100 µM ATP. Inset shows the dependence of the normalized peak current on the K+ concentration. Solid curve is the best fit of the experimental points to Eq. 3.

View larger version (25K): [in this window] [in a new window]   FIGURE 10  Current transients after a 30 mM KCl concentration jump (a), a 100 µM ATP concentration jump (b), and a simultaneous 30 mM KCl and 100 µM ATP concentration jump (c), all in the presence of calcium ions. In all experiments, the same nonactivating solution was used, which contained 250 mM choline chloride, 1 mM MgCl2, 0.225 mM CaCl2, 25 mM TRIS, 0.2 mM EGTA, and 0.2 mM DTT at pH 7.0 (HCl). In experiment a, the activating solution contained 30 mM KCl, 220 mM choline chloride, and the remaining components as in the nonactivating solution; in experiment b, the activating solution had the same composition as the nonactivating solution plus 100 µM ATP; in experiment c, the activating solution differed from that of experiment a by the presence of 100 µM ATP. Inset shows the sole curve a.

View larger version (22K): [in this window] [in a new window]   FIGURE 11  Current transients after 25 mM K+ concentration jumps. Solid curve was obtained with an activating solution containing 25 mM KCl, 125 mM choline chloride, 1 mM MgCl2, 25 mM TRIS, 0.25 mM EGTA, and 0.2 mM DTT at pH 7.0 (HCl), and with a nonactivating solution differing from the activating one by the replacement of 25 mM KCl with 25 mM choline chloride. Dotted curve was obtained under the same conditions as the solid one, apart from the addition of 100 µM TEA chloride to both the activating and the nonactivating solution. Dashed curve was obtained under the same conditions as the solid curve, apart from the addition of 0.6 µM thapsigargin in the cuvette and an incubation period of 10 min before the K+ concentration jump. Inset shows the charge under the solid (1), dotted (2), and dashed curve (3). Error bars express the average standard deviations in the four repeated measurements carried out on the same SSM.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSIONS APPENDIX ACKNOWLEDGEMENTS REFERENCES

ATP concentration jumps
The current transients due to ATP concentration jumps, such as that in Fig. 1, can be fitted with a biexponential function, yielding an ATP-dependent relaxation time constant ₁ and an ATP-independent one, ₂ 300 ms. By performing ATP concentration jumps on fragmented SR adsorbed on a BLM, via the light-induced conversion of caged ATP, Hartung et al. (1997) fitted the resulting current transients with a sum of four exponential functions. The first two relaxation time constants are 5 ms and, therefore, cannot be observed with the technique in this study, because they are shorter than t_peak. The values of the ATP-dependent time constant, denoted by ₃ by the authors, were found to depend somewhat on whether the concentration of the photo-released ATP was changed by varying the caged ATP concentration at constant flash energy or by varying the latter at constant concentration of caged ATP. Values of at saturating ATP were reported to range from 35 to 100 s^–1, depending on the experimental conditions. These values are in fairly good agreement with the value of 50 s^–1 reported in Fig. 2. The dependence of the time constant ₃ upon the ATP concentration was described by a Michaelis-Menten formalism with a half-saturating concentration K_M = 4.6 µM (Hartung et al., 1997), which is close to the value, 2.9 µM, obtained from the fit in the inset of Fig. 1. In view of its dependence on ATP concentration, the relaxation time constant at hand must be related to the binding of ATP to Ca-ATPase. The fourth time constant, ₄, reported by Hartung et al. (1997) is independent of the ATP concentration, is associated with a positive amplitude, and amounts to 330 ms. This time constant accounts for a moderate current overshoot and is entirely analogous to the time constant ₂ reported herein. These authors tentatively ascribe it to the proton countertransport after Ca²⁺ translocation. In principle, however, the overshoot may also be ascribed to the response of the system, consisting of the supporting mixed bilayer and of the adsorbed vesicles, to the pumping of Ca-ATPase, as discussed below.

It is useful to consider our observations in the light of the Ca²⁺-ATPase reaction sequence and a minimal number of partial reactions as outlined in Fig. 12. From the initial linear section of the versus c_peak plot in Fig. 2, a rate constant, k₁, of 1.2 x 10⁷ M^–1s^–1 is obtained. In view of its dependence on ATP concentration, this rate constant must be related, either directly or indirectly, to the binding of ATP to the enzyme. Butscher et al. (1999) reported that phosphorylation and conformational transitions of Ca-ATPase exhibit only minor electrogenicity. It is, therefore, reasonable to conclude that k₁ is to be ascribed to a step after a diffusion-limited ATP binding step in quasi-equilibrium. With this assumption, k₁ is the product of the rate constant for the rate-limiting step and the equilibrium constant for the preceding ATP binding step. Considering an equilibrium constant of 3 x 10⁵ M^–1 for ATP binding (Fig. 1), and 10² s^–1 rate constant for the steps related to enzyme phosphorylation and release of bound Ca²⁺ (Inesi et al., 1988), the resulting product is 3 x 10⁷ M^–1s^–1, in close agreement with the k₁ value obtained in our experiment. It is then apparent that the electrogenicity of the pump is related to lumenal release of Ca²⁺ after enzyme phosphorylation by ATP, as expected.

View larger version (9K): [in this window] [in a new window]   FIGURE 12  Simplified catalytic and transport cycle for the Ca2+-ATPase. Each ATPase molecule has two Ca2+ binding sites and one catalytic site. Phosphorylated enzyme intermediate (E-P) is formed by utilization of ATP, after activation by Ca2+. Interconverting states of the protein display high affinity and cytosolic orientation (E1 and E1-P), or low affinity and lumenal orientation (E2 and E2-P) of the Ca2+ sites. In the forward direction of the cycle, the phosphorylation potential of ATP is utilized to reduce the affinity of the Ca2+ sites. At neutral pH, two Ca2+ are exchanged with two H+. Kinetic and equilibrium constants for the partial reactions were previously characterized in detail (Inesi et al., 1988).

(4)

View larger version (8K): [in this window] [in a new window]   FIGURE 13  Equivalent circuit simulating the mixed bilayer and the SR vesicles adsorbed on it. Ip is the pump current, I the current flowing along the external circuit, and E the external applied potential. Cp and Rp are the capacitance and resistance of the vesicles, and Cm and Rm those of the mixed bilayer.

Equations A2 and Equations A3 show that both I_on and I_off are characterized by a time-independent contribution, –E/(R_m + R_p), which flows along the external circuit during both the activating and inactivating periods. In practice, however, the resistance R_m is so high that this contribution is vanishingly small. This implies that, by our procedure, the potential difference across the vesicular membrane cannot be affected by varying the applied potential E, because any change in E tends to be located across the mixed thiol/lipid bilayer. Equations A2 also shows that, in principle, a finite stationary pump current b generates a stationary contribution to the experimental I_on current; this is given by –[C_m/(C_m + C_p)]b(_c/_m), with _m R_mC_m 1.4 s. In practice, however, _m >> _c, so that the above capacitive coupling eliminates completely the stationary contribution to I_on.

The equivalent circuit adopted herein and represented in Fig. 13 appears to be more realistic for adsorbed flat membrane fragments incorporating integral proteins than for adsorbed proteoliposomes. However, the current is only pumped on the free membrane area, A_f, of the adsorbed vesicles. If we denote by A_c the area of the vesicle-covered surface of the supporting mixed bilayer, approximately identified with the contact area of the vesicles, it can be shown that the experimental on-current, suitably corrected for any current overshoot, is approximately given by I(t) = –{C_m/[C_m(1 + ) + C_p]}I_p(t), with = A_f/A_c (Läuger, 1991). Therefore, the interpretation of experiments with adsorbed membrane sheets and adsorbed vesicles is similar, the only difference being the magnitude of the scaling factor relating I(t) and I_p(t).

Hartung's tentative justification for the overshoot by proton countertransport is disproved by the pH dependence of the current transients due to the ATP-induced Ca²⁺ translocation (see Fig. 7). Thus, the overshoot is more pronounced at pH 8.13, when proton translocation is practically suppressed (see below). Conversely, the explanation of the overshoot by the exponential term of time constant _c in Eq. A2 is supported by the observation that it increases in parallel with a decrease in the amplitude of the off-current (see Fig. 8); Eq. A4 shows that such a decrease is due to a decrease in b, which is expected to determine an increase in the overshoot.

Ca²⁺ concentration jumps in the absence of ATP
The Ca²⁺ dependence of the current transient after a saturating ATP concentration jump is characterized by a half-saturating Ca²⁺ concentration K_1/2 = 1.26 ± 0.08 µM, and by binding constants for the first and second Ca²⁺ ion about equal to K₁ = 1.2 x 10⁶ M^–1 and K₂ = 5.1 x 10⁵ M^–1; these two values denote a moderate cooperativity in the binding. The K_1/2 value is in good agreement with that of 1 µM (range 0.5–1.2 µM), obtained by Hartung et al. (1987) from SR vesicles adsorbed on a BLM. This value is in reasonable agreement with the concentration dependence of the rate of ATP hydrolysis, which is half-maximal at 0.1–0.2 µM Ca²⁺ (Hartung et al., 1987, and references therein).

The current transient after an ATP concentration jump and the charge after a Ca²⁺ concentration jump in the absence of ATP exhibit a very similar dependence upon the Ca²⁺ concentration (compare Figs. 3 and 5). Thus, the half-saturating Ca²⁺ concentration amounts to 1.26 ± 0.08 µM in the first case and to 1.5 ± 0.3 µM in the second. Moreover, both dependencies point to a slight cooperativity in the binding of the two Ca²⁺ ions (an R value of 1.3 in the first case, a Hill coefficient of 1.1 in the second). This result is to be expected, since the magnitude of the current transient after an ATP concentration jump on SR vesicles preincubated in Ca²⁺ is a measure of the amount of Ca²⁺ bound to the pump before its activation. The Ca²⁺ dependence in Fig. 5 is in good agreement with the Ca²⁺ dependence of the increase in tryptophan fluorescence intensity induced by Ca²⁺ binding to Ca-ATPase in the absence of ATP. Thus, the half-saturating Ca²⁺ concentration, K_1/2, in the presence of 1 mM Mg²⁺ amounts to 1.4 µM (see Fig. 6 in Henderson et al., 1994). Somewhat lower values of K_1/2 at pH 6.8–7.4 were reported by Inesi et al. (1980) (0.5 µM) and by Peinelt and Apell (2002) (0.59 µM); moreover, in both cases a Hill coefficient close to 2 was reported, thus suggesting a strong cooperativity in the binding of the two Ca²⁺ ions. The discrepancy between the above high cooperativity and the apparently slight cooperativity found herein can be possibly explained if the binding of the second Ca²⁺ ion to be bound or the release of the second Ca²⁺ ion to be released do not reach full equilibrium conditions during the presteady-state measurements in this study. In fact, a high cooperativity implies that the binding of the second Ca²⁺ ion to be bound, or the release of the second Ca²⁺ ion to be released, is favored with respect to the case that the two binding sites are occupied independently. In this respect it should be noted that strong evidence exists that the binding of the second Ca²⁺ ion is preceded by a conformational change induced by the binding of the first Ca²⁺ ion (Inesi et al., 1980; Henderson et al., 1994), according to the following mechanism:

(5)

Since the phenomenological Hill function ignores the presence of conformational transitions between successive ion binding steps, a Hill coefficient close to unity does not exclude a strong cooperativity. Thus, in the case of the mechanism of Eq. 5, three equilibrium constants must be introduced:

(6)

The fractional saturation y, namely the equilibrium average value of Ca²⁺ ions bound per site, is then given by

(7)

where account has been taken that, in principle, each conformation of the pump may provide two binding sites for Ca²⁺. If K₃ << 1, it may be readily shown that Eq. 7 reduces to

(8)

This equation is identical with Eq. 2, apart from the replacement of the second binding constant K₂ by the product K₂K₃. Therefore, even if the second binding site has a higher affinity for Ca²⁺ than the first (K₂ >> K₁), the "cooperativity parameter" R may still be close to unity when K₃ << 1, namely when the conformational equilibrium is shifted toward the E₁Ca form. From Eq. 8 it is also apparent that the Hill function with a Hill coefficient equal to 2 requires that R be >>1, a situation that may not be satisfied even for K₂ >> K₁, if K₃ << 1. If, in addition, the equilibrium of the conformational step is not fully attained in the current transients in this study, then the net effect is qualitatively analogous to that of an apparently lower K₃ value, and hence of an apparently weaker cooperativity.

Different views on the conformational transition E₁Ca E₁*Ca are reported in the literature. Henderson et al. (1994) regard this transition as in quasi-equilibrium, with an equilibrium constant of unity, on the basis of the belief that tryptophan fluorescence intensity is unlikely to be equally sensitive to the occupancy of the two different binding sites by Ca²⁺ ions. On the other hand, Inesi et al. (1980) consider the conformational transition as a slow step (see also Dupont and Leigh, 1978), appreciably shifted toward the E₁Ca form, on the basis of equilibrium binding data combined with electron paramagnetic resonance spectroscopic measurements on spin-labeled preparations sensitive to conformational changes.

Some evidence for cooperativity in the binding of the two Ca²⁺ ions is provided by the time dependence of the off-current after Ca²⁺ concentration jumps in the absence of ATP. This off-current fits to the sum of two exponentials, whose time constants are reported in Fig. 6. When the Ca²⁺ concentration in the activating solution that precedes the jump of the Ca²⁺-free nonactivating solution is saturating, fast and slow components of the off-current have time constants _1,off 30 ms and _2,off 120 ms. This behavior may be consistent with a sequential mechanism in which Ca²⁺ binding occurs in a protein crevice (Inesi, 1987; Inesi et al., 1990), in which the dissociation of the Ca²⁺ ion that is bound first is blocked by the second Ca²⁺ ion that is bound in the same crevice. Under this assumption, the fast component is due to the release of this second Ca²⁺ ion, whereas the slow component is due to the release of the Ca²⁺ that was bound first. A biexponential decrease in tryptophan fluorescence intensity due to Ca²⁺ dissociation induced by an EGTA concentration jump in the presence of Mg⁺² was reported by Henderson et al. (1994), but the dissociation rate constant for the Ca²⁺ ion released first was considered to be lower than that for the Ca²⁺ ion released second. A biexponential decrease of the intrinsic fluorescence due to Ca²⁺ dissociation from SR vesicles, induced by EGTA in the presence of Mg²⁺, was also reported by Moutin and Dupont (1991); the corresponding relaxation time constants, 18.5 and 102 ms, are relatively close to those obtained herein.

From Figs. 5 and 6 it is apparent that the two time constants _1,off and _2,off increase in parallel with the increase in the amount of Ca²⁺ ions bound to the pump. This behavior cannot be explained by assuming bidirectional Ca²⁺-binding steps in Eq. 5, namely steps in which the backward rate cannot be entirely neglected with respect to the forward one; in fact, the concentration jump of the deactivating solution containing EGTA removes Ca²⁺ ions from the electrode surface. Even upon assuming the presence of residual Ca²⁺ ions in the unstirred layer, the increase in their concentration would cause a decrease in the relaxation time constant, rather than its increase. Thus, in a bidirectional dissociation step PL P + L, where P is the pump, L is the ligand, and ([P] + [PL]) is the constant overall concentration of the pump, the dissociation rate decreases exponentially in time, with a relaxation time constant equal to (k_f + k_b[L])^–1, where k_f and k_b are the rate constants for the forward and backward process. The experimental increase in the two time constants _1,off and _2,off with an increase in the amount of bound Ca²⁺ ions can be explained by assuming that a progressive increase in the number of calcium pumps with two bound Ca²⁺ ions causes a slowdown in their release due to cooperativity in their binding.

The dielectric coefficient of the Ca²⁺ binding steps
From Fig. 4 it is apparent that the charge under the on-current transient due to Ca²⁺ binding to Ca-ATPase in the absence of ATP is less than the charge under the on-current transient due to the ATP-induced Ca²⁺ translocation, as recorded on the same SSM, by a factor of 0.63. To compare these two different charge values, it is first necessary to estimate at which stage of the pump cycle, after its activation by ATP, a stationary pump current is attained. In fact, the capacitive coupling realized by our technique suppresses the measured current as soon as the pump current attains a stationary value (see above). Since the activation of the pump preincubated with Ca²⁺ starts from the E₁*Ca₂ state, it is reasonable to assume that the pump enters the stationary regime immediately after returning to this state. This conclusion is supported by the observation that the charge translocated at pH 7, when full proton countertransport takes place, is practically one-half that translocated at pH 8.2, when proton countertransport is suppressed (see inset in Fig. 7). At pH 7, the sum of the dielectric coefficients of all the steps composing a single turnover of the pump equals 2. Incidentally, the dielectric coefficient of a step is the fraction of the thickness of the membrane, assumed to be a homogeneous dielectric film, across which the charge is translocated during the given step times the translocated charge expressed in electronic units. Let x denote the fraction of the membrane thickness, as measured from the cytoplasmic side, at which the Ca²⁺ binding sites are located in the E₁* conformation. Upon assuming that no appreciable movements of charged residues of the pump take place during ion binding and release, the dielectric coefficients relative to the passage from a group of elementary steps to the subsequent one in a cycle is given by

(9)

in the presence of proton countertransport at pH 7, and by

(10)

in its absence, at pH 8.2. It is readily seen that the only possibility for the charge translocated at pH 8.2 being twice that at pH 7 for any x value is that the stationary pump current starts immediately after the attainment of the next E₁*Ca₂ state. The charges involved in the E₁ E₁*Ca₂ steps and in a whole enzymatic cycle of the Ca pump at pH 7 amount to 50 and 79 pC on the same SSM. Upon regarding them as proportional to 4x and to 2, in view of Eq. 9, an x value of 0.32 is obtained for the fractional distance of the Ca²⁺ binding sites from the cytoplasmic side. This conclusion is in good agreement with the highly resolved three-dimensional structures of the Ca-ATPase of the SR in the E₁ conformation with 2 Ca²⁺ ions bound (Toyoshima et al., 2000) and in the E₂ conformation stabilized by the specific inhibitor thapsigargin (Toyoshima and Nomura, 2002); according to these struct