Interaction of Lanthanide Ions with Bovine Factor X and Their Use in the Affinity Chromatography of the Venom Coagulant Protein of Vipera russeZZi*

SUMMARY The substitution of trivalent Ianthanide ions for Ca(I1) in the Ca(II)-dependent activation of bovine Factor X by the coagulant protein of Russell’s viper venom was studied at pH 6.8. Factor X contains two high affinity metal binding sites which bind Gd(III), Sm(III), and Yb(III) with a & of about 4 X 10P7 M and four to six lower affinity metal binding sites which bind Gd(III), Sm(III), and Yb(II1) with a & of about 1.5 X lo-” M. In comparison, 1 mol of Factor X binds 2 mol of Ca(H) with a & of 3 X 10e4 M and weakly binds many additional Ca(I1) ions. No binding of Gd(II1) to the venom protein was observed. Dy(III), Yb(III), Tb(III), Gd(III), Eu(III), La(III), and Nd(II1) cannot substitute for Ca(I1) in the Ca(II)-dependent


SUMMARY
The substitution of trivalent Ianthanide ions for Ca(I1) in the Ca(II)-dependent activation of bovine Factor X by the coagulant protein of Russell's viper venom was studied at pH 6.8. Factor X contains two high affinity metal binding sites which bind Gd(III), Sm(III), and Yb(III) with a & of about 4 X 10P7 M and four to six lower affinity metal binding sites which bind Gd(III), Sm(III), and Yb(II1) with a & of about 1.5 X lo-" M. In comparison, 1 mol of Factor X binds 2 mol of Ca(H) with a & of 3 X 10e4 M and weakly binds many additional Ca(I1) ions. No binding of Gd(II1) to the venom protein was observed.
Dy(III), Yb(III), Tb(III), Gd(III), Eu(III), La(III), and Nd(II1) cannot substitute for Ca(I1) in the Ca(II)-dependent activation of Factor X by the venom protein at pH 6.8. Kinetic data consistent with the models of competitive inhibition of Ca(I1) by Nd(II1) yielded a Ki of 1 to 4 X 1OV M. The substitution of lanthanide ions for Ca(I1) to promote protein complex formation of Factor X-metal-venom protein without the activation of Factor X facilitated the purification of the coagulant protein from crude venom by affinity chromatography.
Using a column containing Factor X covalently bound to agarose which was equilibrated in 10 mM Nd(III), Tb(III), Gd(III), or La(III), the coagulant protein was purified lo-fold in 40 % yield from crude venom and migrated as a single band on gel electrophoresis in sodium dodecyl sulfate. These data suggest that lanthanide ions compete with Ca(I1) for the metal binding sites of Factor X and facilitate the formation of a nonproductive ternary complex of venom protein-Factor X-metal. Tb (III) fluorescence, with emission maxima at 490 and 545 nm, is enhanced lO,OOO-fold in the presence of Factor X. The study of the participation of an energy donor intrinsic to Factor X in energy transfer to Tb(II1) may be useful in the characterization of the metal binding sites of Factor X. Bovine Factor X, a plasma glycoprotein with a molecular weight of 56,000, participates as a zymogen in an intermediate step during the initiation of blood coagulation (Z-5). Factor X may be activated physiologically by the intrinsic (6) or by the extrinsic pathway (7) of blood coagulation. Alternatively, Factor X may be activated by the coagulant protein of Russell's viper venom (8). Kinetic analyses of this reaction are consistent with the Michaelis-Menten model for enzymes in which the venom protein is an enzyme and Factor X is a substrate (9). The activation of Factor X by the venom protein has an absolute calcium requirement (10). As with many calcium-dependent reactions, the absence of suitable electronic and magnetic properties of calcium has limited the study of the interaction of calcium with Factor X and the venom protein.
For these reasons we have examined the effect of the substitution of lanthanide ions for calcium in the interaction of the venom coagulant protein and Factor X.
In this communication we demonstrate that lanthanide ions bind tightly to the metal binding sites of Factor X, competitively inhibit Ca(II)-dependentFactor X activation by the venom coagulant protein, and facilitate the metal-dependent binding of Factor X and coagulant protein.
We describe an approach for the purification of coagulant protein by affinity chromatography using lanthanide ions to inhibit Ca(II)-dependent catalysis and to facilitate metal-dependent protein complex formation. This method may have general application to the affinity purification of proteins involved in Ca(II)-dependent protein interactions such as those participating in blood coagulation.

AND MATERIALS
Bovine Factor X, purified from fresh bovine plasma by BaSOd adsorption and DEAE-SeDhadex chromatograDhv (22). appeared homogeneous by disc and sodium dodecyl s;fate gel'elkct;obhoresis. Factor X and the coagulant protein of Russell's viper venom activities were assayed as previously described (8). Protein concentration was estimated from the absorbance at 280 nm using an E% nm of 9.5 for Factor X (22) and 13.4 for the venom protein (5). Crude Russell's viper venom (Sigma), 20 mg. was dissolved in 1 ml of 25 mM imidazole, 0.5 M NaCl, pH 6.8,and dialyzed at 4" for 3 hours against 500 ml of the same buffer.
The solution. after clarification by centrifugation at 4,000 x g for 10 min in a Sorvall RC-2B refrigerated centrifuge, was adjusted to 10 mM NdC13 by the addition of 0.1 M NdC13 and incubated for 2 hours at 4". A fine white precipitate that formed was removed by centrifugation.
The supernatant (1 ml) was applied to a column of Sepharose-Factor X (0.7 X 3 cm) equilibrated with 25 mM imidazole, 0.5 M NaCl, 10 mM NdCl,, pH 6.8, at 4". The column was washed with the same buffer at a flow rate of 30 ml per hour until no further protein was eluted, as monitored by absorbance of the eluate at 280 nm. Bound protein was eluted from the column with 25 mM imidazole, 0.5 M NaCl, 10 mM EDTA, pH 6.8, and collected in l-ml fractions.
After the removal of EDTA by exhaustive dialysis against 25 mM imidazole, 0.15 M NaCl, pH 6.8, the fractions were analyzed for protein concentration and coagulant protein activity. When necessary, the venom protein was concentrated in an Amicon ultrafiltrator employing a PM 10 membrane and stored at -15".
Kinetics of Factor X Activation,-Substitution of lanthanide ions for Ca (II) in the activation of Factor X by the venom protein was studied qualitatively using 1 FM, 10 PM, 0.1 mM, or 1.0 mM lanthanide ions in place of 8 mM Ca(I1) in the Factor X assay (8). In other experiments, the kinetics of the Ca(II)-dependent activation of Factor X by the venom protein in the presence of Nd(II1) were studied employing a one-stage assay for activated Factor X (3).
Under the conditions employed, the develooment of activated Factor X from Factor X waslinear for 12.5 min in the presence of 8 mM Ca(I1).
The velocitv of the hvdrolvsis of Factor X by the coagulant.protein is expressed in units of activated Factor X activity generated per min. The reaction, containing 57 rg of Factor X, CaCl,, and NdCl, in 0.3 ml of 25 mM imidazole, pH 6.8 at 37", was initiated with the addition of crude venom (2 pg) in 0.1 ml of 25 mM imidazole, pH 6.8. After 5 or 10 min at 37", a O.l-ml aliquot of the reaction mixture was diluted into 0.4 ml of 15 mM Tris-HCl, 0.1 M NaCl, pH 7.5, at 37", and a 0.1.ml aliquot of this solution was added simultaneously with 0.1 ml of 25 mM CaCh, I5 mM Tris-HCI, 0.1 M NaCI, pH 7.5, to a preincubated mixture of 0.1 ml of pooled human plasma and 0.1 ml of phospholipid at 37". The clotting time was determined and the activated Factor X activity calculated using linear curves constructed from plots of the logarithm of the clotting time (s) versus the logarithm of act.ivated Factor X concentration.
The Ca(I1) and the lanthanide ion concentrations were varied as indicated. Binding of Metal Ions, Factor X, aud Venom Protein-The binding of lanthanide ions and calcium to Factor X or the venom coagulant protein was evaluated at 25", at pH 6.8, by the steady state dialysis method of Colowick and Womack (25) using radioactive lK"Gd(III), i"iSm(III), 169Yb(III), or '%a(II Tri-Carb (model 3390) liauid scintillation spectrometer.
Data were interpreted using al Scatchard plot (27).
In the graphical analysis r is the number of moles of Gd (III) bound per mol of Factor X; c is the molar concentration of unbound Gd(II1).
Linear plots of data describing the upper and lower limits of the slope were obtained by linear regression analysis using a Wang 500 calculator.
Fluorescence spectra were obtained on a Perkin-Elmer model MPF

Interaction of Lanthanide
Ions with Factor X-The binding of lanthanide ions to Factor X was examined by the rate dialysis method (25) using radioactive trivalent lanthanide ions. In experiments with 153Gd(III), the rate of dialysis of 1.6 PM iS3Gd-(111) across the dialysis membrane was 1 x 10L cpm per min in the absence of Factor X and 746 cpm per min in the presence of 19 PM Factor X. The subsequent stcpwise addition of unlabeled GdCL to concentrations ranging from 3.2 PM to 91 PM was associated with a stepwisc increase in the rate of i53Gd(III) dialysis.
At Gd(III) concentrations grcatcr than 25 PM, turbidity of the protein solution was noted in the dialysis cell. Using the analysis of Colowick and Womack (25) to determine the concentration of Gd(II1) free in solution and the concentration of Gd (III) bound to Factor X, these data were interpreted using a Scatchard plot (27)     by rate dialysis at pH 6.8 and 25". These data, analyzed using a Scatchard plot, are presented in Fig. 1B with Factor X-The interaction of calcium (I1) with Factor X was studied at pH 6.8 and 25" by rate dialysis using 45Ca(II) in a solution containing 10 mg per ml of Factor X (1.6 x low4 M).
Although considerable scatter of the data points was noted in multiple experiments, the extrapolation of a line representing the best least mean square fit of the data suggested that 2 mol of Ca(I1) bind to 1 mol of Factor X with a Kd of 3.1 x 1OW M. Additionally, many Ca(I1) ions bind to the protein at higher Ca(I1) concentrations but measurement of these weak interactions was beyond the technical limits of the method.
A summary of the interaction of metals with Factor X is shown in Table I.
Kinetics-Substitution of lanthanide ions for the Ca(I1) ions required for the activat,ion of Factor X by the venom protein was examined.
The presence of Dy(III), Yb(III), Tb(III), Gd- Upper, A column of Sepharose-Factor X was equilibrated at 4" with 10 mM NdCla, 0.5 M NaCl, 25 mM imidazole, pH 6.8; the crude venom-Nd(II1) solution containing 20 mg of protein in 1 ml of buffer was applied and developed with about 15 ml of buffer; bound protein, containing coagulant activity, was eluted using 10 mM EDTA, 0.5 M NaCl, 25 mM imidazole, pH 6.8 (arrow).
Lower, identical experiment as described above except that NdC13 was deleted from the initial equilibration solution.
crude Russell's viper venom in 10 mM NdCl, was applied to a column of Sepharose-Factor X, most of the crude venom protein did not adhere to the derivatized Sepharose (Fig. 3, upper panel). The bound protein, eluted with 10 mM EDTA, exhibited a lofold increase (range was 8-to 15-fold) in the specific activity of the coagulant protein compared to crude venom. About 75% of the original coagulant protein activity applied to the column was recovered in either the bound or the unbound material; one-half of the coagulant protein activity was associated with the bound protein fraction.
Sodium dodecyl sulfate gel electrophoresis of this fraction yielded a major band representing greater than 95% purity and corresponding to a molecular weight of 62,000 (Fig. 4). This is in good agreement with the molecular weight of 60,000 for the venom coagulant protein obtained by sodium dodecyl sulfate gel electrophoresis for protein purified by DEAE-cellulose chromatography and gel filtration on Sephadex G-200 (30). When large quantities of protein were applied to the gels, some low molecular weight material could be identified which was thought to be due to nonspecific binding of protein to the Sepharose-Factor X column.
In control experiments, no protein adhered to the Sepharose-Factor X conjugate in the absence of metal ions ( Fig. 3;  lanthanide solutions in columns at 4'. The use of 1 mM NdCl, yielded smaller quantities of bound protein representing 28% of the applied coagulant protein activity. No protein was bound to the Sepharose-Factor X conjugate in the presence of 0.1 mM NdCla. Affinity chromatography performed with 10 mM NdCl, at 25' consistently resulted in the leaching of bound coagulant protein from the Sepharose-Factor X column prior to elution with EDTA; at 4", recovery of the coagulant protein in the bound fraction was optimized. Maximal binding of the venom coagulant protein was observed when a large excess of crude venom was placed onto the column. When smaller quantities of venom were applied, the protein content of the bound fraction and the specific coagulant protein activity were decreased. Presumably, the binding constant describing the interaction of Factor X and coagulant protein in the presence of Nd(II1) is such that, given the fixed concentration of Factor X bound to the Sepharose, higher ccncentrations of venom coagulant protein increase the amount of Factor X-Nd(III)-venom coagulant protein complex. The specific removal of the venom protein from the Sepharose-Factor X conjugate was facilitated by the chelation of lanthanide ions by A solution containing Factor X (2.8 PM), 0.1 M NaCl, 2.2 PM TbCls at pH 6.8 at 25', was irradiated at 280 nm and the emission spectrum recorded. The slit width of the excitation beam was 16 nm; the slit width of the emission beam was 10 nm. Emission maxima were observed at 490 and 545 nm. The second order scatter peak is centered at 560 nm. The amplitude of the fluorescence emission is given in arbitrary units.
EDTA. EDTA forms very tight complexes with lanthanide ions (11) and competes favorably with the metal binding sites of the protein for the metal ions.
The specificity of the interaction of the venom coagulant protein with Factor X covalently bound to Sepharose was evaluated using columns of Sepharose. In the presence of 10 mM NdCla, no detectable fraction of crude venom bound to the unconjugated Sepharose column which could be eluted with EDTA. These results suggest that the interaction of the venom coagulant protein with Factor X covalently bound to Sepharose is specific and probably simulates the metal-dependent ternary complex formed in solution.

Binding of Terbium(III)
tQ Factor X-The fluorescence properties of Tb(II1) were studied in the presence of Factor X. EXcitation at 280 nm of a solution of Factor X (2.8 PM) and TbCla (2.2 PM) in 0.1 M NaCl at pH 6.8 and 25" produced emission maxima at 490 and 545 nm (Fii. 5) as well as intrinsic tryptophan emission centered at about 344 run. A small maximum at 560 nm represents the second order scatter peak; variation of the excitation wavelength predictably altered the wavelength of this peak. Solutions of Factor X in the absence of Tb(II1) showed intrinsic tryptophan fluorescence but no emission at 490 or 545 nm. The emission spectrum of 2.2 PM Tb(II1) in 0.1 M NaCl at pH 6.8 in the absence of Factor X was not observable while excitation at 280 nm of solutions containing 10 mM Tb(III) in 0.1 M NaCl at pH 6.8 produced emission spectra with maxima which were about one-half of the amplitude of those obtained for 2.2 C(M Tb(II1) in the presence of Factor X. From these data it would appear that Tb(II1) exhibits about a lO,OOO-fold fluorescence enhancement when bound to Factor X. The uncorrected fluorescence excitation spectrum of the Tb(III)-Factor X complex, monitored at 490 run, had a maximum at 283 run; this spectrum was similar to the ultraviolet absorption spectrum of Factor X in the aromatic region. The ultraviolet absorption difference spectrum between Factor X-Tb(II1) uersus Factor X and Tb(II1) was minimal, indicating that the increased Tb(III) fluorescence in the presence of Factor X is not due to an in- in the fluorescence emission spectrum of Tb(III)-Fact,or X. Tb(II1) was added to a solution of Factor X (2.8 PM) and 0.1 mM NaCl, pH 6.8 at 25". The fluorescence emission at 490 nm ( l ---0 ) and 545 nm (A---A) was monitored.
Protein precipitation was noted when the Tb(II1) was added in excess of 25 J.LM.
creased absorption at 280 nm, but due to energy transfer from Factor X to Tb(II1).
These results would suggest that a tyrosine or tryptophan residue in Factor X, in or near a terbium binding site(s), is an energy donor and that the protein-bound Tb (II1) is an energy acceptor. An increase in fluorescence emission at 490 and 545 nm (Fig. 6) and a 10% quenching of intrinsic Factor X fluorescence was associated with the titration of Factor X with Tb(II1) in 0.1 M NaCl at pH 6.8. Turbidity associated with protein precipitation was observed in solutions containing TbCL in excess of 25 PM.
Although the absence of a plateau precluded the complete analysis of the titration experiments, a dissociation constant, Kd, describing the interaction of Tb(II1) and Factor X was roughly estimated to be about 2 x 1OV M per n, where n is the number of Tb(II1) ions bound to 1 mol of Factor X which participates in significant energy transler (31). Because there arc four to six lower affinity metal binding sites determined by the rate dialysis experiments, 72 is an integer between 1 and 6. A Kd between 3 and 19 PM may be estimated from the fiuorescence titration; these values correspond favorably to the dissociation constant determined by rate dialysis describing the interaction of other lanthanide ions with the lower affinity binding sites of Factor X.

DISCUSSION
The mechanism of the activation of Factor X by the venom coagulant protein has been shown to be enzymatic (9), involving proteolytic cleavage of a single bond on Factor X (3-5). The products of this reaction include polypeptide fragments of 44,000 and 11,000 molecular weight (4) whose structures appear to be highly complementary and bind to each other with high affinity (5) The formation of binary and ternary complexes between lanthanide ions and bovine Factor X or the coagulant protein of Russell's viper venom (or both) was investigated with the objective of defining the metal binding properties of Factor X in the presence and absence of the venom protein and the effect of the substitution of lanthanide ions for calcium on the Ca(II)dependent activation of Factor X by the venom protein.
The similarity between the ionic radii and the electrostatic binding of trivalent lanthanide ions and calcium(H) to oxygen ligands originally led to the suggestion that lanthanide ions, with interesting magnetic and electronic properties, might facilitate the physical and biological characterization of Ca(I1) binding proteins (11,12). These ions have subsequently proved useful in characterizing the metal binding sites of Ca(I1) binding proteins (14,16,18,20), in examining three-dimensional structures of proteins and the active site of proteins by x-ray crystallography and nuclear magnetic resonance relaxation techniques (15,16,19,21), and in evaluating the role of metal ions in the catalytic mechanism of hydrolases (13,14,17,18 We have employed kinetic models describing the interaction of the venom protein with Factor X in the presence of Ca(I1) and lanthanide ions. A salient feature of these models is that a single Ca(I1) ion interacts with Factor X and venom protein to facilitate ternary complex formation and Factor X hydrolysis.
Furthermore, trivalent lanthanide ions compete with Ca(II) for the occupancy of this essential metal binding site, and enhance the formation of a stable, nonproductive complex of Factor X-metal-venom protein.
The linearity of the data at each Ca(I1) concentration and the common intercept of all three sets of data are consistent with these models of competitive inhibition. However, because of assumptions in these models, the estimation of the Ki of lanthanide ions in this reaction at pH 6.8 of 1 to 4 PM must be considered a first approximation. This Ki may be compared to the Kd of 0.4 PM and 15 pM describing the interaction of lanthanide ions with the high and lower affinity metal binding sites, respectively, as determined by the rate of dialysis method.
We suggest that, within the experimental uncertainty of this kinetic data, the critical metal binding site(s) which must be occupied by Ca(I1) for activation of Factor X is one (or both) of the high affinity sites on Factor X.
The similarities of certain structural features of the serine proteases make comparison of lanthanide interaction with bovine Factor X and bovine trypsinogen of interest. Trypsinogen and Factor X are zymogens of the serine proteases, trypsin and activated Factor X, respectively, whose active site and NH*terminal amino acid sequences demonstrate marked homology (33,34). Lanthanide ions bind to two metal binding sites on trypsinogen (13)) enhance the rate of trypsin-catalyzed trypsinogen activation (13), and are bound, albeit weakly, to a single metal binding site on porcine trypsin in close proximity of a tryptophan residue (20). It would appear that certain structural features of Factor X and trypsinogen, including metal binding properties, may have been preserved during evolution from a common ancestral protease.
As is the case with transfcrrin (35) and trypsin (20) the lO,OOOfold increase in the intensity of the Tb(II1) emission is due to the energy transfer through a donor in the protein.
Terbium-(III) exhibits a characteristic fluorescence emission spectrum which is due to the j-j electronic transition associated with irradiation by ultraviolet light (36). The magnitude of this emission is enhanced by energy transfer through contact or dipoledipole interactions when Tb(II1) is iiganded in close proximity to a donor fluorophore which can participate as an energy donor. The excitation maximum for Tb(II1) emission of 283 nm for the Factor X-Tb(II1) complex and the association of the quenching of intrinsic tryptophan fluorescence with the binding of Tb-(III) to Factor X suggests that a tryptophan residue may be the energy donor within the protein.
Further studies of energy transfer in the Tb(III)-Factor X interaction should facilitate characterization of the metal binding sites of Factor X. Successful applications of affinity chromatography to the purification of proteins have employed specific ligands covalently bound to an inert matrix and elution systems for the specific removal of proteins which interact with the derivatized matrix. For the purification of enzymes with protein substrates, affinity