Determination of strontium binding to macromolecules

ANALYTICAL

BIOCHEMISTRY

Determination Nam-Won Department

Received

198,

391-393

(1991)

of Strontium

Binding to Macromolecules’

Huh, Pola Berkowitz, Richard G. Hiskey, and Lee G. Pedersen2” of Chemistry,

August

The University

of North

Carolina

at Chapel Hill,

North

Carolina

27514

5, 1991

An equilibrium dialysis technique for determining the binding of strontium to macromolecules is described. The major difficulty to be overcome is that ?3r has a decay product, e”Y, which is also a &emitter. The described protocol is used to determine the Sr binding isotherm to bovine prothrombin fragment 1. The binding is found to be cooperative, somewhat weaker than Ca binding, and to involve approximately nine strontium sites. The stoichiometric equilibrium constants are determined by nonlinear regression. The procedure should be of great utility for many macromolecules that show strontium affinity. o 1991 Academic PAM, I~C.

Strontium is a trace element with a biological importance that is largely unknown. The element is generally regarded to have greatest similarity to calcium in chemical/physiological behavior, primarily because of the studies during the era of above-ground nuclear testing that showed that the P-emitter ?Sr concentrated in high calcium areas in the body. Observed values of “Sr in adult vertebrae in New York City, for instance, reached a maximum value of approximately 2 pCi/g Ca in 1966 and was still half of this value 15 years later (1). Although the ratio of calcium to strontium in the earth’s crust is about 80, the abundance ratio in the human body is 7500 (2). Some suggestions exist that strontium is essential for growth of bone and teeth; however, the evidence is that calcium is preferentially absorbed and strontium preferentially excreted (2). Still, the strontium abundance in human bone is about 360 ppm (2). We have long been interested in the divalent metal ion requirement for the conversion of prothrombin to thrombin. It is generally believed that calcium ion, in ’ This work was supported by Grants HL-27995 (L.G.P.) and HL20161 (R.G.H.) from the National Institutes of Health, U.S. Public Health Service. ’ To whom correspondence should be addressed. ’ Also at National Institute of Environmental Health Sciences, Research Triangle Park, NC 27709 0003-2697/91 Copyright All rights

Chapel Hill,

53.00 0 1991 by Academic Press, of reproduction in any form

the presence of negatively charged phospholipid, factor Xa, and factor Va, is specific for maximal rate of conversion but that strontium will support the reaction at perhaps 20% of the calcium rate (3). Magnesium ion alone will not support thrombin generation; however, in small amounts magnesium has a synergistic effect (4) on the calcium-induced conversion. Prendergast and Mann (4) suggested that the binding of strontium ions to prothombin was weaker than that of calcium ions based on the concentration dependence of the ion-induced fluorescence quench. We have earlier reported equilibrium dialysis results on the binding of calcium (5-7) and magnesium (8) ions to prothrombin fragment 1. While a special protocol was necessary for the magnesium determination due to the short half-life of 28Mg, the calcium determinations, which use 45Ca, are straightforward. Since, to our knowledge, similar determinations have not been conducted for the binding of strontium to specific proteins, we wished to develop a protocol for measuring strontium binding using bovine prothombin as a model system. The study was complicated by the fact that the decay product of ?Sr (‘OY), also is a p emitter and has a short half-life. The essence of this report is a suggestion for overcoming this difficulty and thereby providing a general method for determining the binding isotherm for strontium.

EXPERIMENTAL

PROCEDURES

Materials Proteins. Bovine prothrombin was isolated as described by Mann (9). Bovine prothombin fragment 1 was prepared by incubation of prothombin with ecarin venom, chromatography on DEAE-cellulose, and G-100 chromatography. The protein concentration was determined by uv spectroscopy with the absorbance readings at 280 nm corrected for scattering according to Mann (9). Bovine fragment 1 was studied at a concentration of 20 ELM in each run. A molecular weight of 22,500 (10) and E’% 280 nm = 10.3 was used (9). The protein showed a single band when subjected to sodium dodecyl sulfate-poly391

Inc. reserved.

392

HUH

acrylamide gel electrophoresis on gradient (5-15%) slabs in the buffer described by Laemmli (11) and viewed by Kodavue or Coomassie blue staining. HPLC size-exclusion chromatography with uv detection at 280 nm also showed one peak for the protein preparation. Prior to use, bovine prothombin fragment 1 was stored frozen at -20°C in a solution of 0.1 M NaCl, 0.01 M Tris, 0.02 M sodium azide, pH 7.50, buffer. Reagents. ?Sr (half-life 29 years) obtained from Amersham Corp. (Arlington Heights, IL) was used as a radioactive tracer for Sr(I1) ion. The highest purity commercially available reagents were used throughout. All solutions were prepared from deionized, distilled water (t5 ppb heavy metals, Hydro Inc., Research Triangle Park, NC). Buffers were rendered free from metal ions by passage through a column packed with Chelex100 resin (Bio-Rad). Methods Equilibrium dialysis. Experiments were performed using 10 cells (pH 7.50, 0.10 M NaCl, 0.01 M Tris, 2 ml Teflon cell (Spectrum Medical, Los Angeles, CA)], rotated vertically at 10 rpm for 24 h and thermostatted at 25’C. Fast equilibrating disk membranes (Spectra/Par 2,12-14,000 MWCO, 47 mm) were employed after thorough cleaning (12). Strontium chloride was added to the cells to give final concentrations from 0.05 to 10 mM. Dialysis solutions were handled using Hamilton (Reno, NV) glass syringes with Teflon-tipped plungers and Teflon needles. The standard procedure was to begin with Sr in one of the cells and protein in the other. After equilibration, four aliquots (0.1 ml each) were removed from each cell side and each sample solution was placed in a glass scintillation vial with 0.50 ml EDTA solution and 10 ml of liquid scintillation counting medium (Isolab, Inc.). Samples were counted with a Packard 2000 CD liquid scintillation counter. Determination of thermodynamic binding constants for strontium. The data were analyzed using a Scatchard analysis (S vs V; S = G/[metal free], V = [metal bound]/ [total protein] ) assuming a model with sequential loading of ligands (13) and were fit utilizing the SAS (Statistical Analysis System, Cary, NC). That the binding was near saturation at high r?was verified by a V vs [Sr] free plot (not shown). Determination of Y-90 half-life. It was not possible to obtain an appropriate Scatchard plot using our previous methodology (5-7). We found that after the 24-h equilibration, recovery losses were of the order of 20% and V values were abnormally elevated in the cells with high concentrations of Sr. This difficulty occurred because large amounts of radioactivity was localized in the membranes; this radioactivity was found to decrease rapidly over several days. From the activity of the membranes, we could account for almost all of the recovery

ET

AL. 210000

3

00 0

5

10

15

20

25

30

Days FIG. 1.

The decay curve for the residual radioactivity remaining in washed and dried membranes extracted from the dialysis cells after 24 h of equilibration. The curve has a half-life approximately the same as that of 9.

error. To quantitatively study the effect, we performed a separate equilibrium dialysis experiment in which MWCO 3500 membranes were equilibrated for 24 h. The membrane was then washed with buffer, blot dried, and placed in a scintillation vial as above. The activity was checked daily for 25 days. From the data (Fig. l), The half-life was found to be 65 + 1 h. Since the half-life of 9oY is 64 h (14), this result confirmed that the membrane activity was arising from 9, the immediate product of the decay of ?Sr. The isotope 9oy also decays, in this case to ?Zr, which is stable. Measurement of the ?!3r activity (without 9oY interference). As mentioned above, the Scatchard plots for Sr(I1) binding determined from only 24 h of equilibrium were skewed at high concentrations of Sr(I1). We reasoned that at high Sr(II), some of the @(‘Yinitially bound to the membrane is forced into the protein compartment where it binds to protein with high affinity (the goY ion has a plus 3 charge). While the concentration of bound ‘Oy is small relative to the bound ?Sr, its activity is much greater. A subsequent measurement of the activity to determine Sr(I1) on each side of the membrane is then incorrect. If, however, the scintillation vials are prepared after standard equilibration and then the activity counted every 3 days for 2-3 weeks, the subsequent Scratchard curve becomes well-behaved (Fig. 2). In fact, there is no change in the Scatchard curve determined after 3 weeks. Apparently, after 3 weeks, the goY initially bound to the protein has reverted to the cold BOZr and the steady state activity of goY on each side of the membrane is now approximately equal and cancels in the analysis. The 90Zr occupies only an insignificant fraction of the protein binding sites. An alternate procedure suggested by Rane and Bhatki (15) to separate 9oy from BOSrin the starting hot Sr(I1) solution was attempted. This method is based on the use of Dowex X8 cation exchange resin for separation of 9oY from ?Sr. Although the separation of the ions

Sr BINDING

1000

TO

-

0

2

4

6

6

393

MACROMOLECULES

10

V

FIG.

2. Scatchard plot for strontium ion binding to bovine prothrombin fragment 1. S = c/[metal free], d = [metal bound]/ [protein total]. The solid line represents the fitted binding curve determined using the general sequential binding model.

was uneventful, the subsequent equilibration time needed (we used 12 rather than 24 h to minimize the increase in 9oy due to decay) was such that the 9oY buildup during the equilibration led to similar problems as seen earlier; i.e., the Scatchard plot for the Sr(I1) join binding was skewed to incorrectly large values as high Sr(I1) ion concentrations. This experience tends to confirm our conjecture that one must wait a substantial number of half-lives of 9oy before performing the counting measurements. RESULTS AND DISCUSSION

The Scatchard curve determined for Sr(I1) binding to bovine prothrombin fragment 1 (Fig. 2) shows that the binding is somewhat cooperative with approximately nine sites filled. The equilibrium binding constants found from the nonlinear least squares procedure are K1 = 2075, K, = 2251, K3 = 1311, K4 = 1030, K5 = 740, K6 = 309, K7 = 266, K, = 113, Kg = 10 M-‘. We have previously found this protein to bind approximately seven Ca(I1) ions in a cooperative manner (5,6,7) whereas five Mg(I1) ions are bound noncooperatively. The magnitude of Kl for Ca(I1) binding is 2720 M-‘, quite similar to the Sr(I1) value, whereas Kl is 5900 M-’ for magnesium ions. From these comparisons, we deduce that for binding to prothrombin fragment 1, Sr(I1) ions behave more like Ca(I1) ions than magnesium ions. This conclusion is in keeping with the observation that Mg(I1) ions alone will not support coagulation in a blood clotting assay (4)

whereas Sr(I1) ion will substitute for Ca(I1) ion (3) with about 20% of the Ca(I1) ion efficiency. The results are also in agreement with circular dichroism studies on bovine fragment l/Ca(II), Mg(II), Sr(I1) in which the secondary structure projection from the CD data suggests that the solution structures of the protein with Ca(I1) and Sr(I1) are similar while that with Mg(I1) is somewhat different (16). It is perhaps not unreasonable that the prothrombin fragment 1 when saturated binds five Mg’s, seven Ca’s, and nine Sr’s per molecule since magnesium usually forms hexacoordinate complexes, calcium coordinates from six to eight ligands, and strontium can bind up to nine ligands. The methodology developed in this paper for studying the binding of strontium ions to bovine prothombin fragment 1 is quite general and should be useful for strontium binding studies on other systems. REFERENCES 1. Eisenbud, M. (1987) Environmental demic Press, New York.

Radioactivity,

p. 330, Aca-

2. Hammond, 3.

P. B., and Beliles, R. P. (1980) in Toxicology (Doull, J., Klaassen, C. D., and Amdur, M. O., Eds.), pp. 409-467, Macmillian, New York. Prentice, C. R. M., Breckenridge, R. T., and Ratnoff, 0. D. (1976)

J. Lab. Clin. Med. 69, 229-244. 4. Prendergast, F. G., and Mann, K. G. (1977) J. Biol. Chem. 252, 840-850. 5. Deerfield II, D. W., Berkowitz, P., Olson, D. L., Wells, S., Hoke, R. A., Koehler, K. A., Pedersen, L. G., and Hiskey, R. G. (1986) J. Biol. Chem. 261,4833-4839. 6. Deerfield II, D. W., Olson, D. L., Berkowitz, P., Byrd, P. A., Koehler, K. A., Pedersen, L. G., and Hiskey, R. G. (1987) J. Biol. Chem. 262,4017-4023. I. Deerfield II, D. W., Olson, D. L., Berkowitz, P., Koehler, K. A., Pedersen, L. G., and Hiskey, R. G. (1987) Biochem. Biophys. Res. Commun. 144,520-527. 8. Olson, D. L., Deerfield II, D. W., Berkowitz, P., Hiskey, R. G., and Pedersen, L. G. (1987) Anal. Chem. 160,468-470. 9. Mann, K. G. (1976) in Methods in Enzymology (Lorand, L., Ed.), Vol. 45, pp. 123-156, Academic Press, San Diego. 10. Owen, W. G., Esmon, C. T., and Jackson, C. M.

(1974)

J. Biol.

Chem. 249,594-605. 11. Laemmli, U. K. (1970) Nature 227,680-685. 12. McPhie, P. (1971) in Methods in Enzymology (Jakoby, W. B., Ed.) Vol. 22, pp. 23-32, Academic Press, San Diego. 13. Scatchard, G. (1949) Ann. N.Y. Acod. Sci. 51,660-672. 14. CRC Handbook of Chemistry and Physics (1985) (Weast, R. C., Ed.) 66th ed., p. B268, CRC Press, Boca Roton, FL. 15. Rane, A. T., and Bhatki, K. S. (1966) Anal. Chem. 38,1598-1601. 16. Balbes, L. M., Pedersen, L. G., and Hiskey, R. G. (1991) submitted for publication.