Nd3+ and Co2+ Binding to Sarcoplasmic Reticulum CaATPase AN ESTIMATION OF THE DISTANCE FROM THE ATP BINDING SITE TO THE HIGH-AFFINITY CALCIUM BINDING SITES*

Nd3+ binding to sarcoplasmic reticulum (SR) was detected by inhibition of ATPase activity and directly by a fluorimetric assay. Both methods indicated that Nd3+ inhibited the ATPase activity by binding in the high-affinity Ca2+ binding sites. The stoichiometry of binding was about 11 nmol of Nd3+ bound per mg of SR proteins at pNd = 6.5. At higher [Nd3+], substantial nonspecific binding occurred. The association constant for Nd3+ binding to the high-affinity Ca2+ binding sites was estimated to be near 2 X lo9 M-'. When the CaATPase was inactivated with fluorescein isothiocyanate (FITC), 5.3 nmol were bound per mg of SR protein. This fluorescent probe is known to bind in the ATP binding site. The stoichiometry of Nd3+ binding to FITC-labeled CaATPase was the same, within experimental error, as to the unlabeled CaATPase. Fluorescence energy transfer between FITC in the ATP site and Nd" in the Ca2+ sites was found to be very small. This donor-acceptor pair has a critical distance of 0.93 nm and the distance between the ATP site and the closest Ca2+ was estimated to be greater than 2.1 nm. Parallel measurements with FITC-labeled SR and Co2+, an acceptor with a critical distance 1.2 nm, suggested the ATP and Ca2+ binding sites are greater than 2.6 nm apart. negligibie

Knowledge about the number and nature of and the interactions between several cation binding sites on the CaATPase of SR' is essential in order to understand the enzymatic mechanism of action. There are two high-affinity Caz+ binding sites on the outside of the SR which are converted to lowaffinity sites and exposed to the interior as a result of MgATP binding and hydrolysis. Optimal function of CaATPase requires that K+ and Mg2+ also be bound to the enzyme during at least part of the pumping cycle (1).
Ca2+, M$+, and K+ are not readily detectable by spectro-*This research was supported by National Institute of Health Grants AM25177, GM31083, and AM00509, National Science Foundation Grant CDP-7023045, the California Affiliate of the American Heart Association, and the Pacific Dental Research Foundation. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$ To whom correspondence should be addressed. scopic techniques and an approach used in many laboratories has been to substitute analogs that have better spectroscopic properties, usually for fluorescence or magnetic resonance measurements. Of particular interest here are the members of the lanthanide series, which have been used with SR to substitute for Ca2+ and M$+ and allow convenient spectroscopic determination of association constants or structural properties of the binding sites (2, 3). This group of ions is known to inhibit ATPase activity (4-lo), but experiments can still provide information about the cation binding sites.
Although the Ln3+ cations have ionic radii that are close to that of Ca2+ (11-14), it is becoming clear that they are not always specific analogs for Ca2+ (3, E ) , and that one must identify the type of site to which they bind.
In the experiments reported here, Nd3+ binding to SR vesicles and to FITC-labeled SR was investigated. FITC specifically labels the ATP binding site of the CaATPase, and FITC and Nd3+ are a donor-acceptor pair suitable for fluorescence energy transfer measurements for the distance range of 0.5 to 2 nm. It was found that free Nd3+ at low concentrations inhibits the ATPase activity by specifically binding to the high-affinity Ca2+ binding sites on the CaATPase. The stoichiometry of Nd3+ binding was not changed by labeling with FITC. At higher M ) concentrations of free Nd3+, there was considerable nonspecific binding to the protein and lipids.
Experiments were done at low [Nd3+] using FITC-labeled SR in order to estimate the distance between the Ca2+ and ATP sites. This distance is too great to be measured accurately with this donor-acceptor pair, and is at least 2.1 nm. The distance from FITC to bound Co2+ was estimated to be greater than 2.6 nm, confirming the Nd3+ results. Thus it appears that high-affinity Ca2+ binding sites are not contiguous with the ATP hydrolyzing site. On the contrary, they may be far away from the ATP site.

EXPERIMENTAL PROCEDURES
Proteins and Chemicals-Sarcoplasmic reticulum vesicles were isolated from New Zealand rabbit hindleg muscle by the method of Eletr and Inesi (16). Electrophoresis on polyacrylamide gels in the presence of 1% sodium dodecyl sulfate indicated the vesicles had 80-85% of the protein with M , = 1.1 X lo6, presumably the CaATPase. Typical ATPase activities at 25 "C were 7-8 pmol of Pi min" mg". SR prepared this way forms a maximum of 4-5 nmol of phosphoryated enzyme/mg of protein and binds about 8-10 nmol of Ca2+ per mg of protein (21, 39). Protein concentrations were determined by the biuret method (17), using isolated SR protein ( d r y weight) as a standard. In some cases, relative protein concentrations were determined by intrinsic tryptophan fluorescence intensities. Detergentsolubilized CaATPase monomer was prepared using the nonionic detergent CIzEg as described in Murphy et a!. (31). FITC-labeled CaATPase was prepared by a modification of the method of ; it was found that the Ca2+-activated ATPase activity was abolished when 5.3 nmol of FITC were incorporated per mg of SR protein.

Distance between ATP and Ca2+ Sites on SR CaATPase
Most experiments involving Nd3+ were done in the absence of EGTA. In these cases, SR vesicles were first treated with Chelex 100 by a batch method to remove divalent cations and then Mg2+ and Ca2+ were added to obtain the desired concentrations. Thus, in these cases, [Ca*+] was Ca2' added and low free [Ca"] could not be determined accurately.
Chemicals were of the highest available commercial grade. Chelex 100 was from Bio-Rad. The ionophore A23187 and the detergent C12E9 were from Sigma. Ultrapure Nd203 was from Alfa Chemical Co. and Nd3+ was produced by dissolving the oxide in 6 N HCl and neutralizing with Na02CCH3. Exposure of the Nd3+ solutions to glass was kept to the minimum possible. Tb3+ and La3+ solutions were prepared from ultrapure chloride salts (Alfa).
ATPase Activities-The standard assay conditions were 0.01 mg/ ml of SR protein, 2 mM ATP, 5 mM MgC12, 0.83 mM EGTA, 1 mM CaCI2, 0.3 p M A23187,75 mM KCl, 50 mM MOPS (pH 7.0) at 37 "c. Phosphate production was measured by a phosphomolybdate method (22). Assays that included Nd3+ had 50 PM CaC12 added and no EGTA. When the Ca2+ and M e concentrations were varied, the per cent activity with Nd3+ was corrected for the changes in ATPase activity due to changes in [Ca"] or [ M P ] . Assays that included Co2+ were done with and without EGTA. Free [Co"] was calculated by iteration using a computer. The association constants used were those compiled by Martell and Smith (26).
Spectroscopy-UV absorption spectra and fluorescence spectra were obtained using a MacPherson model EU-700 spectrophotometer and a Perkin-Elmer model MPF-44B fluorospectrophotometer, respectively. In the fluorescence energy transfer experimencs, X,. was 470 nm and X , , was 520 nm. Control measurements with buffer and unlabeled SR indicated that light scattering made a negligibie contribution to the apparent fluorescence in the absence of polarizers or cut off filters.
The quantum yield for FITC-labeled SR was determined by the ratio method using FITC as a standard (Q = 0.85) (23). The overlap integrals for FITC-labeled SR and Nd3+ or Co2+ were determined by a numerical integration using Simpson's composite formula (37) done with a Hewlett-Packard HP-85 calculator. The absorption spectra were those of free Nd3+ and Coz+ because of their low extinction coefficients.
Assavs for Free Nd3+-A modification of the adaptation bv Miller andSeikior (24) of the method of Barela and Sherryi25) for detecting Tb3+ was used to determine the free [Nd3+]. In this method, SR and bound Nd3+ are removed by filtration through Millipore filters or by centrifugation and the filtrate or supernatant is precisely diluted to give about 1 p~ Nd3+ in a solution of 2 p~ Tb3+ and 3.8 pM 2,6dicarboxylpyridine. The decrease in the sensitized Tb3+ fluorescence (Xex25~nm, Xem5450r585 " , ) was compared to a standard curve to obtain the unknown Nd3+ concentration. Linear plots for [Nd3+]'s in narrow ranges centered from around 1 to around 100 p~ Nd3+ were obtainable by varying the [Tb3+] and [2,6-dicarboxylpyridine]. Although Miller and Senkfor (24) recommend Tb3+ to 2,6-dicarboxylpyridine molar ratios of 1:3, for low [Nd3+] the ratio of 1:2 gave more linear responses. An unknown substance in Millipore filters (HA-WP) interfered with this assay, and all the results reported here are from centrifugation experiments.

RESULTS
Inhibition of the ATPase Activity-Nd3+ inhibits the AT-Pase activity of SR. Shown in Fig. 1 is a typical result for the loss of ATPase activity due to increasing added Nd3+. Inhibition of SR ATPase activity by similarly high concentrations of added cation has been reported for many of the Ln3+ ions (3-10, 14) and for Nd3+ in particular (8). If a small excess of EGTA over Nd3+ was added 10 min after the Nd3+, the activity was restored to >90% of its normal value. Increasing the amount of SR in the assay increased the amount of Nd3+ required for a 50% reduction in activity. The averaged results for three [SRI's are shown in the inset of Fig. 1   determined from the intercept and the large stoichiometry determined from the slope were little changed. Experiments are in progress to try to explain this observation.
In order to determine the nature of the inhibitory site, Ca2+ and Mg2+ competition experiments were done. Ca" protected against inhibition by Nd3+, but M$+ did not (Fig. Z), indicating that the Nd3+ inhibits by binding in a Ca2+ binding site. This result parallels that obtained with Tb3+ (3), and taken together with the apparent association constants, show that the Ln3+ inhibition of Ca2+-activated ATPase activity is due to binding in the high-affinity Ca'+ binding sites. The affinity of these sites for Nd3+ obtained from the intercept in Fig. 1 can now be adjusted for the competing free Ca2+ present in the solution, using the equation where KLpp = 1.1 X lo7 M-', Kc. = 4 X lo6 M-', and [Ca'+] = 38 &M. This calculation gives Kapp = 1.7 X lo9 M" for Nd3+ The apparent association constant for SR CaATPase and Tb3+ was inadvertently given as 8.3 X lo6 M" instead of 7.1 X IO6 M" (3).  The effects of Co2+ on the ATPase activity were also investigated. There was enough endogenous Ca2+ to activate the CaATPase to its maximum. Prior treatment of the enzyme with ion-exchange resins or EGTA and gel chromatography did not reduce the activity to the low levels that are obtained with EGTA in the assay. Thus the Co2+-containing experiments had at least micromolar amounts of Ca2+ present. Adding Co2+ instead of, or in addition to, Ca2+ reduced AT-Pase activities to a constant level that was about 38% of the control obtained with only Caz+ (Fig. 3) gives Kapp = 2 X lo9 M" for Co2+ binding to the CaATPase.
The maintenance of constant ATPase activity at [Co"] above 50 p~ and the large Kapp for Co2+ binding suggest that Co2+ may substitute for Caz+ as a substrate for the CaATPase. Preliminary attempts to detect Co2+ uptake were inconclusive and additional work to test this possibility is in progress. It cannot be excluded, at this time, that Co2+ binds to other sites as well as to the high-affinity Ca2+ sites.
Nonspecific Nd3+ Binding-In addition to the specific binding of Nd3+ to sites on the SR vesicles, it is highly probable that nonspecific binding also occurs. The net negative charge of the CaATPase, other proteins, and the lipids, and the high positive charge density of the Nd3+ assure binding will occur even in the presence of millimolar levels of Mg2'. Using a modification of an assay for Ln3+ (24) the free and bound Nd3+ were determined for increasing total [Nd3+] (see "Experimental Procedures"). At [Nd3+] above about M, there is a large and rapidly increasing amount of Nd3+ bound to the vesicles (Fig. 4). The data for measurements made with low [Nd3+] are shown in the inset. All experiments were done in solutions containing 0.1 mg/ml SR, 5 mM MgCI2, 0.1 mM CaC12, 1 mM MOPS, 0.5 M KC1 (pH 7.0), 25 "C.
It is clear from the shape of the curve in Fig. 4 and number of Nd3+ ions bound per mg of SR that substantial nonspecific binding occurs when Nd3+ is present in excess. Even more Nd3+ binds at higher concentrations, but the enzymatic activity was irreversibly lost after 10 min incubation at pNd B 4, and this concentration range was not investigated. At low [Nd3+] (Fig. 4, inset), close to 10 nmol of Nd3+ bind/mg of protein. The average value for Nd3+/SR below pNd = 5.8 was 10.4 k 4 nmol/mg of protein. Table I shows the results for the stoichiometry determined many times at a single low [Nd3+] in solutions containing 10 PM added Ca2+. The result for SR vesicles, 10.6 f 3.6 nmol/mg, is near the value for high-affinity Ca2+ binding sites, and along with the Ca2+ specific reversal of the Nd3+ inhibition of the ATPase activity (Fig. 2), suggests that at low free [Nd3+], specific binding to the high-affinity Ca2+ binding sites is obtained.
Attempts to detect the reversal of the binding of Nd3+ by   3.4 ( n = 9) 11.2 f 5.2 ( n = 12) including high [Ca"] in the buffer were not successful because Ca2+ interferes with the Nd3+ assay (24). In particular, the signal to noise ratio dropped to unusable levels at [Ca"] 5 X M. For this assay, Tb3+ gives a much larger signal than Nd3+ at all Ca2+ concentrations, so Ca2+ reversal of Tb3+ binding was attempted. Tb3+ also inhibits SR ATPase activity (3,8,46). Fig. 5 shows the results for Ca2+ reversal of Tb3+ binding at low [Tb3+] to freshly prepared SR vesicles. Ca2+ appears to reverse the binding of about half the T b 3 + . This is consistent with the observed lack of cooperative Tb3+ binding when it inhibits ATPase activity (3), and suggests that Tb3+ and Ca2+ can simultaneously bind to the CaATPase. However, in ATPase assay buffers, the Ca2+ reversal of Tb3+ inhibition is complete (3, 46) and the Ca2+ binding appears to be cooperative (46). When older vesicles ( 2 3 days) were used, the stoichiometry of Tb3+ binding at low [Ca"] was higher and the Ca2+ reversal was weaker. Nonetheless, the results in Fig.  5 suggest that Tb3+ binding to the high-affinity Ca2+ sites is not cooperative and that half the sites have a very high affinity under equilibrium conditions, without ATP.
FITC Labeling-When vesicular CaATPase was incubated at pH 7.5 with a 15-fold molar excess of FITC, the ATPase activity was lost as the FITC became covalently attached to the enzyme. As shown in Fig. 6, saturation of the attached probe molecules occurred within 30 min (inset) and 5.3 & 0.4 nanomoles of bound FITC/mg of SR protein caused complete inhibition of CaATPase activity (the basal activity was 0.53 Fmol min" mg"). It has been shown that ATP protects against FITC labeling and results indicate FITC binds in the ATP site (20,42). FITC-labeled SR was prepared freshly the day of an experiment by incubating SR vesicles and excess FITC for 30 min in 100 mM KCl, 5 mM MgC12, 10 mM MOPS (pH 7.5, KOH), 100 FM CaC12 at 25 "C, and then removing excess FITC by size exclusion chromatography. In agreement with others (19,20), all the FITC was on the CaATPase, and appeared on the 45-kDa fragment after trypsin treatment when analyzed by gel electrophoresis in the presence of 1% sodium dodecyl sulfate. The stoichiometry of Nd3' binding was not affected by FITC-labeling (Table I).
Fluorescence Energy Transfer-The suitability of fluorescence and Nd3' as a donor-acceptor pair is shown in Fig. 7 for SR labeled with 5.3 nmol of FITC/mg of SR protein (fluorescence spectrum) and free Nd3+ (absorbance spectrum). The overlap integral was 7.23 X cm3 "' when calculated using the equation (29) where F ( X ) and t ( X ) are the fluorescence intensity of the donor and the molar extinction coefficient of the acceptor, respectively, and X is the wavelength. J, and other relevant constants and parameters are listed in Table 11. The critical distance R, is 0.93 nm when calculated from the equation (29) R. , = (50Q0nr-4K2)"6 X 9.7 X 10' nm where Qo is the quantum yield, n, is the refractive index of the medium between the donor and the acceptor, and K 2 is the orientation factor. The approximately spherical symmetry of the acceptor, Nd3+, makes the assumption of K 2 = 2/3 much safer than in cases where both acceptor and donor dipole orientations are unknown (29,30). Given that there is considerable local freedom of motion for the probe FITC attached to SR (38) and the symmetry of the Nd3+, the

Constants for F.E.T. calculations with FZTC, Nd3+, and Co2+
J, was calculated as described in the text, Qo was determined for SR labeled with 5.3 nmol of FITC/mg of SR protein, K 2 is assumed, n, is the refractive index of alanine, R, was calculated as described under "Experimental Procedures."  Efficiencies were determined from decreases in FITC fluorescence intensity at 520 nm. The SR CaATPase was labeled as described in the text. The conditions were as given in Table I Table 111 for vesicular and detergent-solubilized monomeric CaATPase (31). In all cases the efficiency is low, allowing only a minimum distance to be calculated. The relative spatial relationship between the two Ca2+ sites and the ATP site is not known. The sixth power inverse dependency of the efficiency on the distance insures that if the relationship is asymmetric, the nearer Ca2+ site will be the primary acceptor. If the two Ca2+ sites are the same distance from the ATPase, that distance can be obtained. Results for both cases were considered.
For a single donor-acceptor pair, the distance between them can be calculated using the equation (33) where El is the experimental efficiency. E, can also be expressed in terms of the rate constants for excited state energy transfer, fluorescence and dark relaxation, kT, kF, and k D , respectively (33), as If there are n equivalent and equidistant acceptors, then the experimental efficiency in terms of rate constants will be nkT E,, = kD -k kF + flkr and the expressions for El and E, can be combined to give For this multiple acceptor case, the experimental efficiency can be used to calculate an apparent distance, r,, that can be related to the actual distance between the donor and a single acceptor as using the expression above that relates E, and El. In this expression E, is the measured efficiency and rl is the actual distance between the donor and any one of the equivalent and equidistant acceptors. The actual distance, rl, was calculated from the experimental efficiency, E,, using the equation 1 -E,, for n = 1 and 2. Results for fluorescence energy transfer from FITC to Nd3+ and Co2+ are given in Table 111. The rl values are lower limits, but it is clear that the ATP and Ca2+ sites are at least 2 nm and more likely greater than 2.6 nm apart. La3+ is not an acceptor for FITC, and the small change observed for La3+ indicates that the change in fluorescence due to any conformational change in replacing Ca2+ with Ln3+ is small.

CONCLUSIONS
The reversible inhibition of Ca2+-activated ATPase activity by Nd3+ appears to be due to binding to the high-affinity Ca2+ binding sites. The apparent association constant for Nd3+ and the CaATPase, in the presence of ATP, is 2.2 X lo9 "' or greater. This behavior is similar to that observed for the structurally similar Tb3+ (3, 46), and the difference between the two is in semiquantitative agreement with the results of dos Remedios (41).
Nd3+ binding to the high-affinity Ca2+ binding sites also was detected in the absence of ATP, and the stoichiometry was similar to that of Ca2+ binding. The ratio of Nd3+ to FITC was close to 2, in good agreement with that of Ca2+ and FITC (18) or Ca2+ and E -P (39). These data and those for Gd3+ (2) and Tb3+ (3) binding suggest that all the Ln3+ ions will bind to the high-affinity Ca2+ binding sites with association constants >lo9 "'. Thus conditions are known that allow SR vesicles to be specifically labeled at the high-affinity Ca2+ binding sites with Ln3+ cations.
High concentrations of Ca2+ reversed the binding of Nd3+ in the ATPase assay medium, and competed with Tb3+ binding for equilibrium conditions without ATP. The data suggest that Nd3+ and Tb3+ bind strongly to high-affinity Ca2+ binding sites. Labeling with FITC has a negligible effect on the binding of these cations, just as it does on Ca2+ binding (42).
The results for Co2+ are less clear cut. Although it seems likely that Co2+ binds in the high-affinity Ca2+ sites, given the large Kapp and the results for Tb3+ (3) and Nd3+, exclusive binding in the high-affinity sites has not been demonstrated. However, the effect of any additional binding of Co2+ on the efficiency of fluorescence energy transfer would be to increase the efficiency, if the Co2+ were close to the FITC, or not change the efficiency if the Co2+ were far (>3 nm). This makes the distances calculated from the Co2+ data lower limits.
The stoichiometry obtained for FITC labeling confirms the reports that FITC binds preferentially to the ATP binding

Distance between A T P and Ca2+
Sites on SR CaATPase site (18, 20, 27, 28, 42). The binding of Caz+ is unaffected by FITC labeling (41, 42) and the stoichiometry of Nd3+ binding is not changed ( Table I). The ratio of Nd3+ to FITC supports the conclusion that Nd3+ and FITC are binding with high specificity. The fluorescence energy transfer measurements for the CaATPase labeled with FITC and Nd3+ indicate that the high-affinity Ca2+ binding sites are at least 2.1 nm away from the ATP binding site and the Co2+ results suggest they are further than 2.6 nm away (Table 111). Model building of ATP suggests its longest dimension is 1.7 nm. Thus it appears that during Ca2+ transport the action of MgATP binding and hydrolysis in the ATP site must be transmitted a distance through the protein to change the Ca2+ sites. Preliminary reports of steady-state fluorescence and excited state lifetime measurements support this conclusion (43,44), which is consistent with a recent hypothesis that for coupled vectorial transport the ATP site and the Ca2+ sites need to be separated (45).