Tb3+ Binding to Ca2+ and Mg2+ Binding Sites on Sarcoplasmic Reticulum ATPase*

The interactions of Tb3+ and sarcoplasmic reticulum (SR) were investigated by inhibition of Ca2+-activated ATPase activity and enhancement of Tb3+ fluores- cence. Ca2+ protected against Tb3+ inhibition of SR ATPase activity. The apparent association constant for Caz+, determined from the protection, was about 6 x lo6 "l, suggesting that Tb3+ inhibits the ATPase activity by binding to the high affinity Ca2+ binding sites. Mg2+ did not protect in the 2-20 mM range. The association constant for Tb3+ binding to this Ca2+ site was estimated to be about 1 X log M-'. No cooperativity was observed for Tb3+ binding. No enhancement of Tb3+ fluorescence was detected. A second group of binding sites, with weaker affinity for Tb3+, was observed by monitoring the enhancement of Tb3+ fluorescence (LXPBBnm, Lm646nm). The fluorescence intensity increased 950-fold due to binding. Ca2+ did not compete for binding at these sites, but Mg2+ did. The association constant for Mg2+ binding was 94 M-', suggesting that this may be the site that catalyzes phosphorylation of the ATPase by inorganic phos- phate. For vesicles, Tb3+ binding to these Mg2+ sites was best described as binding to two classes of binding sites with negative cooperativity. If the SR ATPase was solubilized in the nonionic detergent CIZEB (dodecyl nonaoxyethylene ether alcohol), in the absence of Ca2+, only one class of Tb3+ binding sites was observed. The total number of sites appeared to remain constant.

did not compete for binding at these sites, but Mg2+ did. The association constant for Mg2+ binding was 94 M-', suggesting that this may be the site that catalyzes phosphorylation of the ATPase by inorganic phosphate.
For vesicles, Tb3+ binding to these Mg2+ sites was best described as binding to two classes of binding sites with negative cooperativity. If the SR ATPase was solubilized in the nonionic detergent CIZEB (dodecyl nonaoxyethylene ether alcohol), in the absence of Ca2+, only one class of Tb3+ binding sites was observed. The total number of sites appeared to remain constant. If Ca2+ was included in the solubilization step, Tb3+ binding to these M g + binding sites displayed positive cooperativity (Hill coefficient, 2.1). In all cases, the apparent association constant for Tb3+, in the presence of 5 mM MgC12, was in the range of 1-5 X lo4 M-'.
The trivalent cations in the lanthanide series are known to inhibit the Ca2+-activated ATPase activity of sarcoplasmic reticulum (1-5). The mechanism of the inhibition is usually assumed to be due to binding of Ln3" to Ca2+ binding sites. The crystal radii of the lanthanide cations are in 96-115-pm Grants AM25177,5207RR05301-13, and AM00509 and National Sci-*This research was supported by National Institutes of Health ence Foundation Grant CDP-7923045. 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.
The abbreviations used are: Ln3+, members of the lanthanide series; SR, sarcoplasmic reticulum; Ca-ATPase, the (Ca2+,Mg2+)-ATPase from rabbit skeletal muscle SR CI2E9, dodecyl nonaoxyethylene ether alcohol (CH3(CH2)11(0CH2CH&OH; Mops, 3-(N-mor-pho1ino)propanesulfonic acid EGTA, ethylene glycol bis(b-aminoether ether)-N,N,N',N'-tetraacetic acid. range, which is similar to that of Ca2+ (99 pm) (6). The Ln3+ ions have been described as Ca2+ analogs (7, 8). Their higher charge density makes them bind more tightly than Ca2+. As an example of their relative affinities, EGTA has about a IO6 times larger association constant for members of the Ln3+ series than it does for Ca2+ (9). Nonetheless, for SR, there is no evidence that Ln3+ inhibition of ATPase activity is due to binding to a Ca2+ binding site. In particular, binding to nonspecific sites or inhibition by TbATP has not been excluded.
Some of the ions in the Ln3+ series have useful spectroscopic properties. Several are paramagnetic and can be used as shift and/or broadening reagents in magnetic resonance measurements (10-12). Others are fluorescent. T b 3 + , Dy3+, and Eu3+ have millisecond excited state lifetimes and are suitable for use in fluorescent energy transfer experiments and as probes of Ca2+ binding sites (13-15). As a beginning of a project to exploit the fluorescent properties of these Ca2+ analogs in studies of SR Ca-ATPase, Tb3+ binding to vesicular and detergent solubilized SR was investigated.
The results indicate that Tb3+ inhibits the ATPase activity of SR by binding to the high affinity Ca2+ binding sites. The affinity of the SR Ca-ATPase for Tb3+ is at least 2 orders of magnitude greater than for Ca2+. T b 3 + bound to these sites is nonfluorescent. Additional Tb3+ binding at lower affinity sites on the enzyme did enhance Tb3+ fluorescence. Surprisingly, these sites are Mg2+ binding sites. Scatchard-type analysis indicated that these M e binding sites interact with one another under some conditions. The nature of the interaction was modified by nonionic detergent solubilization and depended on the [Ca"] at the time of solubilization.

EXPERIMENTAL PROCEDURES
Proteins and Chemicals-Sarcoplasmic reticulum vesicles were isolated from New Zealand rabbit hind leg muscle by the method of Eletr and Inesi (16). Electrophoresis on polyacrylamide gels in the presence of 1% sodium dodecyl sulfate indicated that the vesicles had 85% of the protein with a M, = 1.1 X lo5, presumably the (Ca2+,M$+)-ATPase. Typical ATPase activities were 7-8 pmol of Pi min" mg". SR prepared this way forms a maximum of about 4 nmol of phosphorylated enzyme/mg of protein and transports about 8 nmol of ea"/ mg of enzyme. Protein concentrations were determined by the biuret method (17), using bovine serum albumin as a standard without correction. In some cases, relative protein concentrations were determined by intrinsic tryptophan fluorescence intensities.
Experiments involving Tb3+ were done in the absence of EGTA.

Fluorescent Tb3+ Binding to SQ
ClzEg were from Sigma. Ultrapure Tbc13 was from Alfa Chemical Co. and kept under argon until used. All Tb3+ solutions were kept in plastic or quartz containers. ATPase Activities-The standard assay conditions were 0.01 mg/ ml of SR protein, 5 mM MgC12, 0.83 mM EGTA, 1 mM CaC12, 0.8 nM A23187,75 mM KCl, 50 mM MOPS-KOH, pH 7.0, at 37 "C. Phosphate production was measured by a phosphomolybdate method (19). Assays that included T b 3 + had no EGTA and 50 p~ CaClz added. Assays done in the presence of C12E9 usually did not contain A23187. The affinity of ATP for Tb3+ in the assay buffer (minus A23187 and SR) was determined from the decrease in ATP absorbance at 260 nm due to Tb3' binding.
Fluorescence Measurements-All measurements were made on a Perkin-Elmer MPF-44B fluorospectrophotometer equipped with a DCSU-2 microprocessor. Tryptophan fluorescence was measured at 330 nm from solutions irradiated at 295 nm. Tb3+ fluorescence was measured at 545 nm from solutions irradiated a t 295 nm. A 430 nm cut-off filter was used to minimize any contribution from light scattering. The temperature was 25 f 0.1 "C. A titration took from 10 to 20 min.

RESULTS AND DISCUSSION
ATPase Inhibition-Tb3+ inhibits SR vesicle ATPase activity. The loss of activity as a function of added [Tb3+] is shown in Fig. 1 for [SR] between 0.01 and 0.1 mg/ml. The free [Tb3+] is much lower due to chelation by ATP. Also shown, in the inset, is the [Tb3+] required for 50% inhibition uersus the [SR]. One point is included from the original observation of Tb3+ inhibition of SR Ca-ATPase activity by dos Remedios (3). The inhibition could be completely reversed by including enough EGTA in the assay to chelate the T b 3 + , but not the Ca2+. No T b 3 + fluorescence enhancement was detected for solutions that were completely inhibited. To eliminate the large inner filter effect due to ATP, measurements of T b 3 + fluorescence (hex295nmr &m545nm) were also made on solutions without ATP containing added [Tb3+] greater than the estimated free [Tb3+] required for complete inhibition. No increase was observed, indicating that binding to the inhibitory site does not enhance Tb3+ fluorescence significantly.
There was no protection against inhibition of the ATPase activity when the [M?] was varied between 2.5 and 15 mM ( Fig. 2A). Ca2+, on the other hand, protected against inhibition. The ratio of ATPase activity without and with Tb3+ increases and approaches 100% as the [Ca'+] is increased (Fig.  2B). The [Ca"] that reverses 50% of the inhibition is about 40 X 1O"j M, which corresponds to an apparent K ' , near 3 X inhibition is due to binding in a Ca2+ binding site. The lack of an effect due to Mg+ eliminates TbATP as the inhibiting species.

The magnitude of the apparent association constant (K'I)
for Tb3+ binding to the inhibiting site in the presence of Ca2+ was estimated from free [ T b 3 + ] values calculated assuming that ATP is the only significant chelator of T b 3 + . This is a reasonable assumption, since the assay conditions were 0.01-0.2 mg/ml of SR, 50 p M CaCk 5 mM MgCl,, 75 mM KCl, 50 mM MOPS-KOH, 0.83 nM A23187, and 2 mM ATP. The association constant for ATP and T b 3 + in the assay buffer was determined to be 1.4 X lo5 M" (see "Experimental Procedures"), in reasonable agreement with the value of 0.9 x IO5 M" obtained for the similar ion (La3+) binding to ATP under similar conditions (2). The apparent free [Tb3+] that gives 50% inhibition of the ATPase activity still increases with [SR] (Fig. 3)  that K'I is an apparent association constant, and perhaps a lower limit.
The data for T b 3 + inhibition and Ca2+ protection can be shown to be consistent with Tb3+ binding in the high affinity Ca2+ binding site. If one assumes it is the high affinity site Enhancement of Tb3+ Fluorescence-At higher [Tb3+], fluorescence peaks were observed at 490, 545, 580, and 640 nm in the presence of SR vesicles irradiated at 295 nm. This type of behavior is expected if Tb3+ binds near enough to an aromatic amino acid for fluorescence energy transfer from the amino acid to the Tb3+ to occur. Equimolar concentrations of vesicles composed of SR phospholipids had a much smaller fluorescence enhancement ((2%). The fluorescence intensities at 545 nm for SR vesicles and SR lipid vesicles are plotted in Fig. 5 as a function of added Tb3+. The conditions (5 mM MgC12, 10 mM MOPS-KOH, pH 7.0) were chosen to minimize Tb3+ binding to phospholipids or to soluble anions. Using the association constant determined below, the fluorescence of Tb3+ was estimated to increase 950-fold. If 1 mM EGTA was added to a solution of T b 3 + and SR, the fluorescence was reduced to that of the EGTA in the absence of SR within a few minutes. The lack of La3+ penetration to the interior of SR vesicles (5), the absence of exposed proteins other than the Ca-ATPase on the vesicle exterior (23), and the rapid reversal of Tb3+ binding in the presence of EGTA all suggest that Tb3+ fluorescence enhancement is due to binding to site(s) on the portion of the Ca-ATPase on the exterior of the vesicles. These sites are in addition to the high affinity Ca2+ sites that are saturated by T b 3 + at much lower ITb3+].
A Scatchard-type analysis (24) of the fluorescence enhancement due to binding to the vesicles indicated there were at least two classes of sites (Fig. 6). The downward curvature is compatible with independent sites or with negatively cooperative sites. The points obtained with 5 m M MgClz present can be fit adequately, assuming two classes of sites of equal number with association constants of 3.2 X lo4 M" and 0.65 Also shown in Fig. 6 is the fact that Ca2+ (up to 1 mM) did not affect Tb3+ binding. Surprisingly, since Tb3+ is often considered to be a Ca2+ analog, M$+ was an effective inhibitor of Tb3+ binding to the sites that enhance fluorescence at 545 nm (Fig. 6). Using the [Tb3+] for 50% increase in fluorescence to estimate average apparent association constants for Tb3+ in the presence ( K & ) and absence (Kam) of M2+, the association constant for Mg2+ as a competitive inhibitor was determined from the equation KMg = 94 f 9 M-'. This value for KMg is near that for Mf binding in the sites that catalyze phosphorylation of the Ca-ATPase by inorganic phosphate (26-28), suggesting that Tb3+ may be binding to those sites in this case.
Detergent Solubilization-Dispersion of the SR Ca-ATPase in micelles of the nonionic detergent ClzEg (29) did not affect the net total fluorescence enhancement, suggesting that the number of bound Tb3+ ions did not change. The background fluorescence for SR lipid vesicles in ClzEg was increased, indicating some interaction between the Tb3+ and the detergent. However, since the [Tb3+] for 50% of the increase due to binding SR protein in C12E9 was only a few per cent less than it is for the vesicle suspension, Tb3+ binding to either phospholipids or to the detergent appears to be stoichiometrically negligible compared to the free Tb3+. Purification of the Ca-ATPase.detergent complex by size exclusion chromatography did not change the magnitude of the fluorescence enhancement, but required times that lead to inactivation of the Ca-ATPase under some of the conditions used, so titrations were done immediately after solubilization.
The results for a Tb3+ titration of SR solubilized in &E9 were found to depend upon the presence or absence of Ca2+ at the time of solubilization. In either case, the SR vesicles were first treated with Chelex 100 (Na+) to remove all divalent cations. Next, 10 mM MOPS-KOH, pH 7.0,5 mM MgClz, and 0 or 0.1 mM CaClz were added. Then C12E9 in the respective buffer was added to obtain 20 mg/ml of detergent. The solution was then titrated with Tb3+. The total net fluorescence enhancement at 545 nm was the same for solubilization with or without Ca2+ present, but the results obtained by Scatchard analysis were different. When Ca2+ was omitted from the solubilization step, there was a single class of sites ( Fig. 7) with an association constant of (4.9 f 1.9) X lo4 M-', close to the value for the higher affinity site determined for vesicles at the same [MgZ+]. Adding Ca2+ up to 1 mM after solubilization without Ca2+ and incubating up to 30 min had no effect. M$+ inhibited Tb3+ binding with about the same apparent association constant as observed for the vesicle case (Fig. 7B). It seems that without Ca2+ present, C,,Eg converts all the M$+ binding sites to a single type that appears similar with regard to affinity for Tb3+ to the higher affinity class of sites on the vesicles. If 0.1 mM Ca2+ was included in the solubilization, Scatchard analysis indicated positive cooperativity for Tb3+ binding to the M e binding sites (Fig. 8). M g + still inhibited Tb3+ binding. The apparent affinity constant for Tb3+ in the presence of 5 mM MgC12 is 2.0 X lo4 "' . A Hill coefficient nH = 2.1) was estimated by averaging the data for all [Mf]. Thus, there are at least two sites involved. The positive cooperativity observed under these conditions suggests that the Scatchard curves with downward curvature, observed for Tb3+ binding to vesicles, reflect negative cooperativity between the sites. Ca2+ seems to be required at the time of solubilization to maintain communication between the sites. The loss of these site-site interactions is apparently irreversible in detergent.
The ATPase activity for SR solubilized with or without Ca2+ present was at least 85% of that of vesicles after times sufficient to complete a titration. Longer times are known to lead to inactivation (29).

CONCLUSIONS
The sites at which Tb3+ binding causes inhibition of the SR ATPase activity appear to be the high affinity Ca2+ binding sites. The quantitative agreement between the affinity for Ca2+ binding determined from the protection experiments and the well established value for Ca2+ binding to activate ATPase activity, the lack of protection by M e , and the apparent absence of significant additional binding of Tb3+ to the vesicles all support this conclusion. The association constant for Tb3+ binding to these sites is about lo9 M-'. This is about lo3 times greater than Ca2+ binding and corresponds to around 16 kJ/mol greater standard free energy. This is considerably smaller than the increase of about 33 kJ/mol observed for T b 3 + and Ca2+ binding to EGTA and demonstrates the high specificity for Ca2+ of the high affinity Ca2+ binding sites on SR.
There was no indication of cooperativity for Tb3' binding in sites where Ca2+ binding involves positive cooperativity (30-32). Gd3+ binding to these high affinity sites (11) seemed Fluorescent Tb3+ Binding to Sarcoplasmic Reticulum ATPase to involve negative cooperativity or independent sites with different affinities. Taken altogether, these results suggest that the interaction between the high affinity Ca2+ binding sites depends on the nature of the cation bound and is capable of wide variation. The binding sites that enhance T b 3 + fluorescence are clearly M e binding sites. Tb3+ is often thought to be a Ca2+ analog, because of their similar crystal radii. Mg2+ is much smaller, and the results shown here indicate that T b 3 + and probably all the lanthanide tervalent cations are not specific for Ca2+ binding sites.
Site-site interactions between the M e binding sites were observed for Tb3+ binding. The nature of this interaction is curious because it shows striking changes, depending on the conditions. Positive, negative, and no cooperativity were observed for Tb3+ binding when the [Ca"] and lipid composition were varied. Ca2+ seemed to be essential for the cooperative site-site interactions to occur. Loss of the cooperative interactions between the Ca2+ sites, due to Ca2+ deprivation, has also been observed for SR vesicles (33); but that loss was reversible. In the present case, Ca2+ appears to act as a link of communication between the two M e binding sites. Once removed in the presence of detergent, the link does not seem to be re-established. When the linkage is present, the lipid environment appears to modify the nature of the interaction between the sites. The natural phospholipids promote negative cooperativity, and the detergent C12E9 promotes positive cooperativity, at least for Tb3+ binding.