ARCHIVES
OF BIOCHEMISTRY
AND
BIOPHYSICS
Vol. 288, No. 1, July, pp. 185-191, 1991
The Binding of Terbium Ions to Gelsolin Reveals Two Classes of Metal Ion Binding Sites Ross L. Tellam’ Department of Experimental Patholog.y, John Curtin School of Medical Research, Australian Natioial University, Canberra A.C.T. 2601, Au&r&a
Received October 10, 1990, and in revised form February
7, 1991
Spectroscopically active terbium ions have been used to probe the Ca2+ ion-binding sites on human plasma gelsolin. The luminescence of Tb3+ ions bound to gelsolin is markedly enhanced when excited indirectly at 295 nm due to Forster type dipole-dipole energy transfer from neighboring tryptophan residues. Titration of this luminescence with increasing concentrations of Tb3+ ions was saturable although the shape of this titration curve was complex indicating the involvement of multiple classes of sites. Luminescence lifetime measurements (obtained by indirect excitation at 295 nm) demonstrate the presence of two classes of sites characterized by a major lifetime of 1.0-l. 1 ms and a minor lifetime of 0.70.8 ms. However, while the amplitude of the minor lifetime showed a hyperbolic dependence on the Tb3+ ion concentration, the amplitude of the major lifetime showed a strongly sigmoidal dependence. Different classes of Tb3+ ion binding sites can also be distinguished by the different Ca2+ ion concentrations needed to displace Tb3+ ions from these sites on gelsolin. It is proposed that the occupancy of one class of Tb3+ ion binding sites on gelsolin causes a conformational change in gelsolin which then allows a second class of cryptic Tb3+ ion binding sites to be expressed. The implications of these results in terms of the binding of Ca2+ ions to gelsolin and the regulation of o 1991 the activities of gelsolin by calcium are discussed. Academic
Press,
Inc.
Gelsolin is a calcium-regulated actin-binding protein which has marked effects on the formation of actin filaments in vitro and is present in a wide range of mammalian tissues (l-7). Gelsolin severs actin filaments, nucleates (or accelerates) the assembly of actin polymer from ’ To whom correspondence should be addressed at CSIRO Division Tropical Animal Production, CSIRO Private Mail Bag 3, Indooroopilly, QLD 4068, Australia. Fax (617)8707034.
oow9861/91
$3.00
Copyright 0 1991 by Academic Press, Inc. All rights of reproduction in any form reserved.
actin monomer and caps the normally fast growing end of an actin filament (1,4,8). The severing and nucleating activities are Ca2+ ion-dependent in the range of the free intracellular calcium ion concentration (0.1-l PM) (1, 5). In view of the activities of gelsolin, it has been suggested that this protein may be important in the microfilament reorganization which accompanies the activation of various cells by extracellular agents (4, 9). Many of these agents induce intracellular calcium ion fluxes. Central to the understanding of the functions of gelsolin is a requirement for detailed knowledge of the binding of Ca2+ ions to this protein. Direct 45Ca2+binding measurements have concluded that cytoplasmic and plasma gelsolins’ have approximately two independent and equivalent Ca”+ ion binding sites (& = 0.6-1.0 PM; Refs. (10-12)). However, there is now accumulating evidence that the binding of Cazt ions to gelsolin is not as simple as was first thought. First, the stoichiometries measured in the direct 45Ca2+binding studies are actually significantly less than two. Indeed, the stoichiometry measured for the plasma gelsolin isoform is only 1.4-1.6 (12). Second, the Ca2+ ion concentration dependence of the change in viscosity of actin polymer solutions caused by cytoplasmic gelsolin occurs over a relatively narrow range of Ca2+ ion concentrations (0.15-0.4 PM; 13) which is inconsistent with a simple model involving two independent and equivalent Ca2+ ion binding sites with a Kd = 1 PM. Third, a Ca2+ ion-dependent conformational change has been demonstrated in both isoforms of gelsolin (14-18). Analysis of the Ca2+ ion concentration dependence of this conformational change gives an apparent Ca2+ ion binding constant of 20 PM for plasma gelsolin and a Hill coefficient of 2 (14). This binding constant is approximately 20- to ‘Plasma gelsolin is nearly identical to cytoplasmic gelsolin except that the former has an amino terminal extension of 27 amino acids which is presumably required to act as a signal sequence to facilitate secretion of the plasma gelsolin isoform (22).
185
186
ROSS L. TELLAM
30-fold greater than that obtained by direct 45Ca2+binding studies with plasma gelsolin. Moreover, the Hill coefficient of 2 suggests that there is cooperativity in the binding of Ca2+ ions to gelsolin-a result in contrast with the direct 45Ca2t ion binding studies. Fourth, the majority of studies demonstrating a Ca” ion dependence of the activities of gelsolin were carried out at Ca2+ ion concentrations much greater than the apparent Ca2+ ion binding constant (& = 1 PM). This raises the possibility that weaker, but functionally important Ca2+ ion binding sites may also be involved in the expression of the activities of gelsolin. Indeed, Doi et al. (19) recently have presented evidence for the presence of high (& = 7 PM) and low (& = 1 mM) affinity Ca2+ ion binding sites on plasma gelsolin. In view of these inconsistencies, the spectroscopically active metal ion, terbium, has been used in the current study to probe the calcium ion binding sites on gelsolin. Both the Tb3+ and Ca2+ ions have similar ionic radii, similar coordination numbers, and a preference for binding to charged or uncharged oxygen groups (20,21). Although terbium is not always an isomorphic replacement for calcium, Tb3+ ions are able to substitute for Ca2’ ions on a large number of Ca2+ ion-binding proteins and in many instances support the Ca2+-dependent biological activities of these proteins (21). The present study reports the differentiation of Tb3+ ion-binding sites on gelsolin into two distinct classes. The implications of this result in terms of the binding of Ca2+ ions to gelsolin and the expression of the Ca2+ ion-dependent activities of gelsolin are discussed. EXPERIMENTAL
PROCEDURES
Chemicals. Terbium (III) chloride hexahydram was purchased from Aldrich Chemical Co., Inc. Dipicolinic acid was obtained from Sigma Chemical Co. (St. Louis, MO). All other reagents were of analytical grade and distilled-deionized water was used for all solutions. Plasticware was prewashed in EGTA3 and then rinsed copiously with distilleddeionized water before use to minimize calcium contamination. Atomic absorption spectroscopy demonstrated that there was less than 0.4 @M contaminating Ca*’ ions present in the solutions used in this study. Gelsolin activity assay. The ability of gelsolin to nucleate the assembly of actin polymer from actin monomer (i.e. to increase the initial rate of bivalent cation-induced actin polymerization) has been used to provide a rapid and highly sensitive means for determining the activity and purity of gelsolin (4). Briefly, this assay consisted of final concentrations of 5 1M rabbit skeletal muscle actin monomer and 0.1 pM pyreneeactin which were equilibrated to ‘20°C in 2 mM Tris, 0.2 mM CaCl,, 0.2 mM ATP, 1.5 mM NaN3, pH 8.0 (G-buffer). Actin polymerization was initiated by making the solution 3 mM in MgClz. The subsequent increase in the fluorescence intensity (ex., 365 nm; em., 407 nm) of the pyreneactin as it is incorporated into actin polymer was monitored and used as a measure of the rate of actin polymerization (4). A small volume (~10% of the total assay volume) of the sample to be tested for nucleating
3 Abbreviations used: EGTA, ethylene glycol bis(P-aminoethyl ether) N,N’-tetraacetic acid; PMSF, phenylmethylsulfonyl fluoride; DTT, dithiothreitol; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis.
activity was added simultaneously with the addition of the MgCl*. The ionic conditions of the polymerization were chosen to overcome any contributions which may arise from the variable ionic components of the sample and in addition, to accentuate the differences in the initial rate of polymerization in the absence and presence of gelsolin. There was a linear relationship between the concentration of purified gelsolin (up to 0.3 pM) and the initial rate enhancement factor (initial rate in the presence of sample/initial rate in the absence of sample). Gelsolin was isolated and purified from human Gelsolin purification. plasma obtained from the Australian Red Cross Blood Bank (Woden, ACT). The method used for the purification of gelsolin was based on a procedure suggested by Dr. J. Bryan. Three methods were used to localize the gelsolin during the purification: measurement of gelsolin nucleating activity; enrichment of a 93,000 A4, band on SDS-PAGE; and immunoblotting with an affinity purified, rabbit anti-human platelet gelsolin antibody. The purification was carried out at 4’C. Briefly, human plasma (containing 2 mM PMSF) was subjected to ammonium sulfate fractionation (30-50%), dialyzed against a buffer containing 20 mM Tris-HCI, 0.2 mM DTT, 2 mM EGTA, 80 mM NaCl, and 2 mM PMSF, pH 7.8 (Ebuffer), and added to a DEAE-Sepharose column equilibrated with the same buffer. After washing the column with 1 bed volume of this buffer, gelsolin was eluted with a linear 80-200 mM NaCl gradient in the same buffer. The active fractions were then concentrated by ammonium sulfate precipitation (60%), dialyzed against 20 mM Tris-HCl, 1 mM CaClz, 1 mM PMSF, pH 7.8 (C-buffer), and added to a DEAE-Sepharose column equilibrated with the same buffer. Gelsolin was eluted by a linear O-200 mM NaCl gradient (in C-buffer) and was well separated from the major peak of eluted protein. The central feature of this purification scheme is a decrease in the binding affinity of gelsolin to DEAE-Sepharose in the presence of Ca2+ ions. This gelsolin was 93% pure as judged by densitometry of a 5-fig sample subjected to SDSPAGE. Further chromatography on Sephacryl S-200 was performed on some preparations to improve their purity. The purified gelsolin was stored at -70°C in E-buffer plus 3 mg sucrose/mg gelsolin. The protein maintained its activity for at least 2 months under these conditions. Before use, the gelsolin was extensively dialyzed against 0.1 M NaCl, 1 mM EGTA, and 0.1 a buffer containing 20 mM Tris-HCl, mM PMSF, pH 7.3, and then against a similar buffer but in the absence of EGTA and PMSF. The effect of Tb3+ ions on gelsolin function could not be determined because actin also binds Tb3+ ions. The concentration of gelDetermination of protein concentrations. solin was determined spectrophotometrically at 280 nm using an extinction coefficient of 1.63 liter. g-’ (14) and a molecular weight of 83,000 (22). Skeletal muscle actin was purified from rabbits (23) and its concentration determined spectrophotometrically at 290 nm using an extinction coefficient of 0.63 liter * g-’ (24) and a molecular weight of 42 300 (25). Luminescence spectroscopy. Fluorescence and long-lived fluorescence (luminescence) measurements were performed in a Perkin-Elmer LS-5 luminescence spectrometer equipped with a thermostated cell holder set at 2O“C. Spectra are uncorrected for the variation in detector efficiency with wavelength although the instrument does contain an internal rhodamine B reference standard that corrects for lamp intensity fluctuations. Luminescence spectra were measured with excitation and emission monochromator slit widths of 5 nm. The instrument contains a pulsed xenon arc lamp which provides a band of energy with a halfwidth at peak intensity of less than 10 ps. The gating of the sample monochromator is delayed by a time, td, to eliminate the effects of the flash. In general, td was 0.1 ms or greater. Consequently, there was no contribution from fluorescence or light scattering in the emission signal. A gating time (tp, i.e., the time period after td) was chosen for the acquisition of data. For titrations and spectra, maximal sensitivity was achieved with ta = 0.1 ms and t, = 9.99 ms. Excitation was at 295 nm (indirect excitation) or 460 nm (direct excitation) and luminescence emission was measured at 554 nm for all titrations. The 295 nm excitation wavelength was chosen to minimize primary self-absorption effects of
BINDING
OF TERBIUM
187
TO GELSOLIN
(b) 8-
6-
650
550
600
Wavelength
500
300
400 Wavelength
(rim)
(nm)
FIG. 1. Luminescence spectra of Tb3’ in the presence of gelsolin. Gelsolin (4.5 pM) in 20 mM Tris-HCl, 100 mM NaCl, pH 7.3, was mixed with 200 pM TbC& and the resulting luminescence emission spectrum (a) recorded with indirect excitation at 295 nm and the corresponding excitation spectrum (b) measured at an emission wavelength of 554 nm (continuous lines). The comparable spectra in the absence of gelsolin are represented by the dashed lines. Other conditions were slit widths 5 nm each, td = 0.1 ms, t, = 9.9 ms, and the temperature was maintained at 20°C.
the protein and to selectively excite tryptophan residues (26). Titrations were performed by adding small volumes of a stock concentration of TbC13 to 500 ~1 of the sample. The total change in volume was always less than 5%. There was no significant Tb3+ ion precipitation at pH 7.3 as measured by the lack of change in the luminescence intensity of a terbium solution (using direct excitation) after high-speed centrifugation (lOO,OOOg,60 min). Luminescence lifetimes were measured by setting t, = 0.1 ms and increasing td from 0.1 to 9.99 ms in small increments. The resulting luminescence decay data were fitted, by a standard least squares fitting routine (27), to a two-component exponential decay function; P(t) = Al[exp(ptdlT1)l
+ Uexp(-UT,)],
[II
where P(t) is the luminescence intensity at time td, A, and A, are the amplitudes of the two components, and T, and T2 are the luminescence lifetimes of the two components. In addition, the nonlinear curve-fitting routine (Marquardt-Levenberg algorithm) available on the commercial software package Sigma Plot (Jandel Scientific) was also used. The fit of the data to a two-component exponential decay process is defined in this software package by the magnitude of the parameter dependencies. Values greater than 0.99 indicate that the data have been over-modeled. The parameter dependencies for AL, AZ, T,, T2 were all less than 0.95 indicating that the model was appropriate for the data. Resolution of the data into three or more components was not able to be reliably performed. The luminescence lifetime of the Tb3+-dipicolinic acid complex was determined (ex., 455 nm; em. 554 nm) to ascertain the accuracy of the lifetime measurements. This data fitted well to a single exponential decay with a calculated lifetime of 2.11 f 0.03 ms which is identical to a previously reported value (28). RESULTS
The luminescence of Tb3+ ions has been used extensively to probe the Ca2+ ion binding sites on a wide range
of proteins (21). The emission spectrum of Tb3+ consists of four well defined peaks in the region between 460 and 660 nm. The most intense of these peaks (-550 nm) is the result of a 5D4+7F5 transition which gives a longlived luminescence emission. There are two pathways leading to the excited 5D4 state, by direct and indirect excitation. In the former case, all of the protein-bound Tb3+ ions can be equally excited at 460-480 nm. However, the low molar absorptivity (extinction coefficient = 0.05 M-‘) of bound Tb3’ ions in this wavelength region results in a very weak emission signal (20, 29) and consequently gives results which are difficult to analyze quantitatively because of their uncertainty. The second and much more efficient method, by indirect excitation, results from resonance energy transfer from aromatic amino acids in proteins (excited at 270-295 nm) to neighboring proteinbound Tb3+ ions (21, 29). Figure la shows the luminescence emission spectra for Tb3+ ions alone and Tb3+ ions in the presence of gelsolin after excitation at 295 nm. There was negligible signal arising from gelsolin in the absence of terbium (not shown). The Tb3+ ion emission is markedly enhanced in the presence of gelsolin. The uncorrected emission spectrum shows well defined peaks at 498, 554, and 595 nm and a shoulder at 630 nm. The peak at 554 nm is the most intense. (The results shown in Fig. 2 demonstrate that gelsolin is saturated with Tb3+ ions under these conditions.) The corresponding luminescence excitation spectra of these solutions are shown in Fig. lb. The excitation spectrum for Tb3+ ions alone shows characteristic
188
ROSS L. TELLAM I
I
1
1
300
400
60
0 0
200
100 TbCl
3
Concentratmn
(PM)
FIG. 2. The Tb3’ ion concentration dependence of the luminescence intensity obtained by indirect excitation of Tb”+ ions. Gelsolin (150 nM) was titrated with increasing concentrations of TbCl, and the Tb3+ ion luminescence intensity (at 554 run), obtained by indirect excitation at 295 nm, was recorded for each concentration. (O), in the presence of gelsolin; (+), in the absence of gelsolin. Other conditions were the same as those described in the caption to Fig. 1 except that the excitation slit width was 15 nm and the emission slit width was 20 nm. A small background signal (- 1.5 units) in the absence of Tb”’ ions was subtracted from all measurements.
major peaks at 460 and 230-nm (not shown) and in between these wavelengths there are a number of minor peaks. In the presence of gelsolin the intensity of the excitation peak at 460 nm is increased by 35%; however in addition a new, major peak with a maximum at 278 nm has appeared. This wavelength region is characteristic of the absorption spectrum of proteins and in particular, gelsolin (14). These results demonstrate that resonance energy transfer occurs from donor aromatic amino acids in gelsolin to adjacent gelsolin-bound Tb3+ ions. The excitation spectra of a number of Tb3+ ion binding proteins have been characterized (26). The ratio of emission intensities with excitation at 292 and 276 nm can be used as a diagnostic criterion for the involvement of particular types of aromatic amino acids. The calculated value of this ratio from Fig. lb is 0.63 which is in the 0.5-1.0 range, defining the principal involvement of tryptophan residues. Consistent with the evidence for resonance energy transfer between tryptophans and bound Tb3’ ions is the small quenching (-10%) of the intrinsic fluorescence of gelsolin in the presence of saturating concentrations of Tb3+ ions (result not shown). The dependence of the Tb3+ ion luminescence intensity (ex. at 295 nm) on the Tb3+ ion concentration in the presence and absence of gelsolin is shown in Fig. 2. The signal derived from Tb3+ ions in the absence of gelsolin was directly proportional to Tb 3+ ion concentration and relatively small in magnitude compared with the signal in the presence of gelsolin. The luminescence signal in the presence of gelsolin is saturable at Tb”+ ion concentrations
in excess of 150 PM, thereby demonstrating that there are specific metal ion binding sites on gelsolin. The nonhyperbolic nature of this titration is consistent with at least two or more classes of gelsolin-bound Tb”+ ions. The first class of sites is occupied initially at the lower Tb3+ ion concentrations (O-20 FM). As a consequence of the occupancy of this class of sites, a second class of sites is subsequently filled. This second class of sites is associated with a much greater luminescence enhancement compared with the first class of sites. The effect of Ca2+ ions on the luminescence of a solution of gelsolin saturated with Tb”+ ions is shown in Fig. 3. Sufficiently high Ca2+ ion concentrations (>250 mM) reduce the gelsolin-enhanced Tb3+ ion luminescence to a level commensurate with that for the same concentration of free Tb3+ ions. This result suggests that Ca2+ ions compete with Tb”+ ions for the same binding sites on gelsolin. There are two distinct Ca2’ ion concentration ranges that cause a decrease in the luminescence of this solution of gelsolin and Tb3+ ions. The first, at a Ca”+ ion concentration of approximately l-5 PM gives about a 14% reduction in the luminescence signal while the second range occurs at much higher Ca2+ ion concentrations (-100 mM) and results in the near complete inhibition of the signal. This latter effect was not due to an increased ionic strength as an additional 200 mM NaCl had little effect on the enhanced signal. The enhanced luminescence obtained by indirect excitation (i.e., at 295 nm) of a solution of gelsolin and Tb3+ ions allows sufficient signal sensitivity to measure the luminescence decay (and therefore luminescence lifetime,
-6
-4
0
-2 log
[CaC12]
(M)
FIG. 3. Effect of Ca’+ ions on the indirect luminescence intensity of gelsolin saturated with TV+. Gelsolin (200 nM) was saturated with 200 /.tM Tb3+ and the resulting luminescence was measured after indirect excitation at 295 nm. The corresponding signal in the absence of gelsolin was approximately 5 units for this concentration of Tb3+ ions. Other conditions were the same as those described in the caption to Fig. 2.
BINDING
OF TERBIUM
nitudes 1.0-1.1 ms and 0.7-0.8 ms throughout most of this range of Tb3’ ion concentrations. At low Tb”+ ion concentrations the decay is represented by a single lifetime equal to 0.8 ms. The amplitude (A,) of the greater lifetime component (T,) shows a strongly sigmoidal dependence on Tb 3+ ion concentration while that for the minor component (A,) is approximately hyperbolic. Indeed, there is virtually complete resolution of these two components at low Tb3+ ion concentrations (
OF BIOCHEMISTRY
AND
BIOPHYSICS
Vol. 288, No. 1, July, pp. 185-191, 1991
The Binding of Terbium Ions to Gelsolin Reveals Two Classes of Metal Ion Binding Sites Ross L. Tellam’ Department of Experimental Patholog.y, John Curtin School of Medical Research, Australian Natioial University, Canberra A.C.T. 2601, Au&r&a
Received October 10, 1990, and in revised form February
7, 1991
Spectroscopically active terbium ions have been used to probe the Ca2+ ion-binding sites on human plasma gelsolin. The luminescence of Tb3+ ions bound to gelsolin is markedly enhanced when excited indirectly at 295 nm due to Forster type dipole-dipole energy transfer from neighboring tryptophan residues. Titration of this luminescence with increasing concentrations of Tb3+ ions was saturable although the shape of this titration curve was complex indicating the involvement of multiple classes of sites. Luminescence lifetime measurements (obtained by indirect excitation at 295 nm) demonstrate the presence of two classes of sites characterized by a major lifetime of 1.0-l. 1 ms and a minor lifetime of 0.70.8 ms. However, while the amplitude of the minor lifetime showed a hyperbolic dependence on the Tb3+ ion concentration, the amplitude of the major lifetime showed a strongly sigmoidal dependence. Different classes of Tb3+ ion binding sites can also be distinguished by the different Ca2+ ion concentrations needed to displace Tb3+ ions from these sites on gelsolin. It is proposed that the occupancy of one class of Tb3+ ion binding sites on gelsolin causes a conformational change in gelsolin which then allows a second class of cryptic Tb3+ ion binding sites to be expressed. The implications of these results in terms of the binding of Ca2+ ions to gelsolin and the regulation of o 1991 the activities of gelsolin by calcium are discussed. Academic
Press,
Inc.
Gelsolin is a calcium-regulated actin-binding protein which has marked effects on the formation of actin filaments in vitro and is present in a wide range of mammalian tissues (l-7). Gelsolin severs actin filaments, nucleates (or accelerates) the assembly of actin polymer from ’ To whom correspondence should be addressed at CSIRO Division Tropical Animal Production, CSIRO Private Mail Bag 3, Indooroopilly, QLD 4068, Australia. Fax (617)8707034.
oow9861/91
$3.00
Copyright 0 1991 by Academic Press, Inc. All rights of reproduction in any form reserved.
actin monomer and caps the normally fast growing end of an actin filament (1,4,8). The severing and nucleating activities are Ca2+ ion-dependent in the range of the free intracellular calcium ion concentration (0.1-l PM) (1, 5). In view of the activities of gelsolin, it has been suggested that this protein may be important in the microfilament reorganization which accompanies the activation of various cells by extracellular agents (4, 9). Many of these agents induce intracellular calcium ion fluxes. Central to the understanding of the functions of gelsolin is a requirement for detailed knowledge of the binding of Ca2+ ions to this protein. Direct 45Ca2+binding measurements have concluded that cytoplasmic and plasma gelsolins’ have approximately two independent and equivalent Ca”+ ion binding sites (& = 0.6-1.0 PM; Refs. (10-12)). However, there is now accumulating evidence that the binding of Cazt ions to gelsolin is not as simple as was first thought. First, the stoichiometries measured in the direct 45Ca2+binding studies are actually significantly less than two. Indeed, the stoichiometry measured for the plasma gelsolin isoform is only 1.4-1.6 (12). Second, the Ca2+ ion concentration dependence of the change in viscosity of actin polymer solutions caused by cytoplasmic gelsolin occurs over a relatively narrow range of Ca2+ ion concentrations (0.15-0.4 PM; 13) which is inconsistent with a simple model involving two independent and equivalent Ca2+ ion binding sites with a Kd = 1 PM. Third, a Ca2+ ion-dependent conformational change has been demonstrated in both isoforms of gelsolin (14-18). Analysis of the Ca2+ ion concentration dependence of this conformational change gives an apparent Ca2+ ion binding constant of 20 PM for plasma gelsolin and a Hill coefficient of 2 (14). This binding constant is approximately 20- to ‘Plasma gelsolin is nearly identical to cytoplasmic gelsolin except that the former has an amino terminal extension of 27 amino acids which is presumably required to act as a signal sequence to facilitate secretion of the plasma gelsolin isoform (22).
185
186
ROSS L. TELLAM
30-fold greater than that obtained by direct 45Ca2+binding studies with plasma gelsolin. Moreover, the Hill coefficient of 2 suggests that there is cooperativity in the binding of Ca2+ ions to gelsolin-a result in contrast with the direct 45Ca2t ion binding studies. Fourth, the majority of studies demonstrating a Ca” ion dependence of the activities of gelsolin were carried out at Ca2+ ion concentrations much greater than the apparent Ca2+ ion binding constant (& = 1 PM). This raises the possibility that weaker, but functionally important Ca2+ ion binding sites may also be involved in the expression of the activities of gelsolin. Indeed, Doi et al. (19) recently have presented evidence for the presence of high (& = 7 PM) and low (& = 1 mM) affinity Ca2+ ion binding sites on plasma gelsolin. In view of these inconsistencies, the spectroscopically active metal ion, terbium, has been used in the current study to probe the calcium ion binding sites on gelsolin. Both the Tb3+ and Ca2+ ions have similar ionic radii, similar coordination numbers, and a preference for binding to charged or uncharged oxygen groups (20,21). Although terbium is not always an isomorphic replacement for calcium, Tb3+ ions are able to substitute for Ca2’ ions on a large number of Ca2+ ion-binding proteins and in many instances support the Ca2+-dependent biological activities of these proteins (21). The present study reports the differentiation of Tb3+ ion-binding sites on gelsolin into two distinct classes. The implications of this result in terms of the binding of Ca2+ ions to gelsolin and the expression of the Ca2+ ion-dependent activities of gelsolin are discussed. EXPERIMENTAL
PROCEDURES
Chemicals. Terbium (III) chloride hexahydram was purchased from Aldrich Chemical Co., Inc. Dipicolinic acid was obtained from Sigma Chemical Co. (St. Louis, MO). All other reagents were of analytical grade and distilled-deionized water was used for all solutions. Plasticware was prewashed in EGTA3 and then rinsed copiously with distilleddeionized water before use to minimize calcium contamination. Atomic absorption spectroscopy demonstrated that there was less than 0.4 @M contaminating Ca*’ ions present in the solutions used in this study. Gelsolin activity assay. The ability of gelsolin to nucleate the assembly of actin polymer from actin monomer (i.e. to increase the initial rate of bivalent cation-induced actin polymerization) has been used to provide a rapid and highly sensitive means for determining the activity and purity of gelsolin (4). Briefly, this assay consisted of final concentrations of 5 1M rabbit skeletal muscle actin monomer and 0.1 pM pyreneeactin which were equilibrated to ‘20°C in 2 mM Tris, 0.2 mM CaCl,, 0.2 mM ATP, 1.5 mM NaN3, pH 8.0 (G-buffer). Actin polymerization was initiated by making the solution 3 mM in MgClz. The subsequent increase in the fluorescence intensity (ex., 365 nm; em., 407 nm) of the pyreneactin as it is incorporated into actin polymer was monitored and used as a measure of the rate of actin polymerization (4). A small volume (~10% of the total assay volume) of the sample to be tested for nucleating
3 Abbreviations used: EGTA, ethylene glycol bis(P-aminoethyl ether) N,N’-tetraacetic acid; PMSF, phenylmethylsulfonyl fluoride; DTT, dithiothreitol; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis.
activity was added simultaneously with the addition of the MgCl*. The ionic conditions of the polymerization were chosen to overcome any contributions which may arise from the variable ionic components of the sample and in addition, to accentuate the differences in the initial rate of polymerization in the absence and presence of gelsolin. There was a linear relationship between the concentration of purified gelsolin (up to 0.3 pM) and the initial rate enhancement factor (initial rate in the presence of sample/initial rate in the absence of sample). Gelsolin was isolated and purified from human Gelsolin purification. plasma obtained from the Australian Red Cross Blood Bank (Woden, ACT). The method used for the purification of gelsolin was based on a procedure suggested by Dr. J. Bryan. Three methods were used to localize the gelsolin during the purification: measurement of gelsolin nucleating activity; enrichment of a 93,000 A4, band on SDS-PAGE; and immunoblotting with an affinity purified, rabbit anti-human platelet gelsolin antibody. The purification was carried out at 4’C. Briefly, human plasma (containing 2 mM PMSF) was subjected to ammonium sulfate fractionation (30-50%), dialyzed against a buffer containing 20 mM Tris-HCI, 0.2 mM DTT, 2 mM EGTA, 80 mM NaCl, and 2 mM PMSF, pH 7.8 (Ebuffer), and added to a DEAE-Sepharose column equilibrated with the same buffer. After washing the column with 1 bed volume of this buffer, gelsolin was eluted with a linear 80-200 mM NaCl gradient in the same buffer. The active fractions were then concentrated by ammonium sulfate precipitation (60%), dialyzed against 20 mM Tris-HCl, 1 mM CaClz, 1 mM PMSF, pH 7.8 (C-buffer), and added to a DEAE-Sepharose column equilibrated with the same buffer. Gelsolin was eluted by a linear O-200 mM NaCl gradient (in C-buffer) and was well separated from the major peak of eluted protein. The central feature of this purification scheme is a decrease in the binding affinity of gelsolin to DEAE-Sepharose in the presence of Ca2+ ions. This gelsolin was 93% pure as judged by densitometry of a 5-fig sample subjected to SDSPAGE. Further chromatography on Sephacryl S-200 was performed on some preparations to improve their purity. The purified gelsolin was stored at -70°C in E-buffer plus 3 mg sucrose/mg gelsolin. The protein maintained its activity for at least 2 months under these conditions. Before use, the gelsolin was extensively dialyzed against 0.1 M NaCl, 1 mM EGTA, and 0.1 a buffer containing 20 mM Tris-HCl, mM PMSF, pH 7.3, and then against a similar buffer but in the absence of EGTA and PMSF. The effect of Tb3+ ions on gelsolin function could not be determined because actin also binds Tb3+ ions. The concentration of gelDetermination of protein concentrations. solin was determined spectrophotometrically at 280 nm using an extinction coefficient of 1.63 liter. g-’ (14) and a molecular weight of 83,000 (22). Skeletal muscle actin was purified from rabbits (23) and its concentration determined spectrophotometrically at 290 nm using an extinction coefficient of 0.63 liter * g-’ (24) and a molecular weight of 42 300 (25). Luminescence spectroscopy. Fluorescence and long-lived fluorescence (luminescence) measurements were performed in a Perkin-Elmer LS-5 luminescence spectrometer equipped with a thermostated cell holder set at 2O“C. Spectra are uncorrected for the variation in detector efficiency with wavelength although the instrument does contain an internal rhodamine B reference standard that corrects for lamp intensity fluctuations. Luminescence spectra were measured with excitation and emission monochromator slit widths of 5 nm. The instrument contains a pulsed xenon arc lamp which provides a band of energy with a halfwidth at peak intensity of less than 10 ps. The gating of the sample monochromator is delayed by a time, td, to eliminate the effects of the flash. In general, td was 0.1 ms or greater. Consequently, there was no contribution from fluorescence or light scattering in the emission signal. A gating time (tp, i.e., the time period after td) was chosen for the acquisition of data. For titrations and spectra, maximal sensitivity was achieved with ta = 0.1 ms and t, = 9.99 ms. Excitation was at 295 nm (indirect excitation) or 460 nm (direct excitation) and luminescence emission was measured at 554 nm for all titrations. The 295 nm excitation wavelength was chosen to minimize primary self-absorption effects of
BINDING
OF TERBIUM
187
TO GELSOLIN
(b) 8-
6-
650
550
600
Wavelength
500
300
400 Wavelength
(rim)
(nm)
FIG. 1. Luminescence spectra of Tb3’ in the presence of gelsolin. Gelsolin (4.5 pM) in 20 mM Tris-HCl, 100 mM NaCl, pH 7.3, was mixed with 200 pM TbC& and the resulting luminescence emission spectrum (a) recorded with indirect excitation at 295 nm and the corresponding excitation spectrum (b) measured at an emission wavelength of 554 nm (continuous lines). The comparable spectra in the absence of gelsolin are represented by the dashed lines. Other conditions were slit widths 5 nm each, td = 0.1 ms, t, = 9.9 ms, and the temperature was maintained at 20°C.
the protein and to selectively excite tryptophan residues (26). Titrations were performed by adding small volumes of a stock concentration of TbC13 to 500 ~1 of the sample. The total change in volume was always less than 5%. There was no significant Tb3+ ion precipitation at pH 7.3 as measured by the lack of change in the luminescence intensity of a terbium solution (using direct excitation) after high-speed centrifugation (lOO,OOOg,60 min). Luminescence lifetimes were measured by setting t, = 0.1 ms and increasing td from 0.1 to 9.99 ms in small increments. The resulting luminescence decay data were fitted, by a standard least squares fitting routine (27), to a two-component exponential decay function; P(t) = Al[exp(ptdlT1)l
+ Uexp(-UT,)],
[II
where P(t) is the luminescence intensity at time td, A, and A, are the amplitudes of the two components, and T, and T2 are the luminescence lifetimes of the two components. In addition, the nonlinear curve-fitting routine (Marquardt-Levenberg algorithm) available on the commercial software package Sigma Plot (Jandel Scientific) was also used. The fit of the data to a two-component exponential decay process is defined in this software package by the magnitude of the parameter dependencies. Values greater than 0.99 indicate that the data have been over-modeled. The parameter dependencies for AL, AZ, T,, T2 were all less than 0.95 indicating that the model was appropriate for the data. Resolution of the data into three or more components was not able to be reliably performed. The luminescence lifetime of the Tb3+-dipicolinic acid complex was determined (ex., 455 nm; em. 554 nm) to ascertain the accuracy of the lifetime measurements. This data fitted well to a single exponential decay with a calculated lifetime of 2.11 f 0.03 ms which is identical to a previously reported value (28). RESULTS
The luminescence of Tb3+ ions has been used extensively to probe the Ca2+ ion binding sites on a wide range
of proteins (21). The emission spectrum of Tb3+ consists of four well defined peaks in the region between 460 and 660 nm. The most intense of these peaks (-550 nm) is the result of a 5D4+7F5 transition which gives a longlived luminescence emission. There are two pathways leading to the excited 5D4 state, by direct and indirect excitation. In the former case, all of the protein-bound Tb3+ ions can be equally excited at 460-480 nm. However, the low molar absorptivity (extinction coefficient = 0.05 M-‘) of bound Tb3’ ions in this wavelength region results in a very weak emission signal (20, 29) and consequently gives results which are difficult to analyze quantitatively because of their uncertainty. The second and much more efficient method, by indirect excitation, results from resonance energy transfer from aromatic amino acids in proteins (excited at 270-295 nm) to neighboring proteinbound Tb3+ ions (21, 29). Figure la shows the luminescence emission spectra for Tb3+ ions alone and Tb3+ ions in the presence of gelsolin after excitation at 295 nm. There was negligible signal arising from gelsolin in the absence of terbium (not shown). The Tb3+ ion emission is markedly enhanced in the presence of gelsolin. The uncorrected emission spectrum shows well defined peaks at 498, 554, and 595 nm and a shoulder at 630 nm. The peak at 554 nm is the most intense. (The results shown in Fig. 2 demonstrate that gelsolin is saturated with Tb3+ ions under these conditions.) The corresponding luminescence excitation spectra of these solutions are shown in Fig. lb. The excitation spectrum for Tb3+ ions alone shows characteristic
188
ROSS L. TELLAM I
I
1
1
300
400
60
0 0
200
100 TbCl
3
Concentratmn
(PM)
FIG. 2. The Tb3’ ion concentration dependence of the luminescence intensity obtained by indirect excitation of Tb”+ ions. Gelsolin (150 nM) was titrated with increasing concentrations of TbCl, and the Tb3+ ion luminescence intensity (at 554 run), obtained by indirect excitation at 295 nm, was recorded for each concentration. (O), in the presence of gelsolin; (+), in the absence of gelsolin. Other conditions were the same as those described in the caption to Fig. 1 except that the excitation slit width was 15 nm and the emission slit width was 20 nm. A small background signal (- 1.5 units) in the absence of Tb”’ ions was subtracted from all measurements.
major peaks at 460 and 230-nm (not shown) and in between these wavelengths there are a number of minor peaks. In the presence of gelsolin the intensity of the excitation peak at 460 nm is increased by 35%; however in addition a new, major peak with a maximum at 278 nm has appeared. This wavelength region is characteristic of the absorption spectrum of proteins and in particular, gelsolin (14). These results demonstrate that resonance energy transfer occurs from donor aromatic amino acids in gelsolin to adjacent gelsolin-bound Tb3+ ions. The excitation spectra of a number of Tb3+ ion binding proteins have been characterized (26). The ratio of emission intensities with excitation at 292 and 276 nm can be used as a diagnostic criterion for the involvement of particular types of aromatic amino acids. The calculated value of this ratio from Fig. lb is 0.63 which is in the 0.5-1.0 range, defining the principal involvement of tryptophan residues. Consistent with the evidence for resonance energy transfer between tryptophans and bound Tb3’ ions is the small quenching (-10%) of the intrinsic fluorescence of gelsolin in the presence of saturating concentrations of Tb3+ ions (result not shown). The dependence of the Tb3+ ion luminescence intensity (ex. at 295 nm) on the Tb3+ ion concentration in the presence and absence of gelsolin is shown in Fig. 2. The signal derived from Tb3+ ions in the absence of gelsolin was directly proportional to Tb 3+ ion concentration and relatively small in magnitude compared with the signal in the presence of gelsolin. The luminescence signal in the presence of gelsolin is saturable at Tb”+ ion concentrations
in excess of 150 PM, thereby demonstrating that there are specific metal ion binding sites on gelsolin. The nonhyperbolic nature of this titration is consistent with at least two or more classes of gelsolin-bound Tb”+ ions. The first class of sites is occupied initially at the lower Tb3+ ion concentrations (O-20 FM). As a consequence of the occupancy of this class of sites, a second class of sites is subsequently filled. This second class of sites is associated with a much greater luminescence enhancement compared with the first class of sites. The effect of Ca2+ ions on the luminescence of a solution of gelsolin saturated with Tb”+ ions is shown in Fig. 3. Sufficiently high Ca2+ ion concentrations (>250 mM) reduce the gelsolin-enhanced Tb3+ ion luminescence to a level commensurate with that for the same concentration of free Tb3+ ions. This result suggests that Ca2+ ions compete with Tb”+ ions for the same binding sites on gelsolin. There are two distinct Ca2’ ion concentration ranges that cause a decrease in the luminescence of this solution of gelsolin and Tb3+ ions. The first, at a Ca”+ ion concentration of approximately l-5 PM gives about a 14% reduction in the luminescence signal while the second range occurs at much higher Ca2+ ion concentrations (-100 mM) and results in the near complete inhibition of the signal. This latter effect was not due to an increased ionic strength as an additional 200 mM NaCl had little effect on the enhanced signal. The enhanced luminescence obtained by indirect excitation (i.e., at 295 nm) of a solution of gelsolin and Tb3+ ions allows sufficient signal sensitivity to measure the luminescence decay (and therefore luminescence lifetime,
-6
-4
0
-2 log
[CaC12]
(M)
FIG. 3. Effect of Ca’+ ions on the indirect luminescence intensity of gelsolin saturated with TV+. Gelsolin (200 nM) was saturated with 200 /.tM Tb3+ and the resulting luminescence was measured after indirect excitation at 295 nm. The corresponding signal in the absence of gelsolin was approximately 5 units for this concentration of Tb3+ ions. Other conditions were the same as those described in the caption to Fig. 2.
BINDING
OF TERBIUM
nitudes 1.0-1.1 ms and 0.7-0.8 ms throughout most of this range of Tb3’ ion concentrations. At low Tb”+ ion concentrations the decay is represented by a single lifetime equal to 0.8 ms. The amplitude (A,) of the greater lifetime component (T,) shows a strongly sigmoidal dependence on Tb 3+ ion concentration while that for the minor component (A,) is approximately hyperbolic. Indeed, there is virtually complete resolution of these two components at low Tb3+ ion concentrations (
