ANALYTICAL
BIOCHEMISTRY
199,
293-299
( 1991)
Determination of the Affinity of Vitamin D Metabolites to Serum Vitamin D Binding Protein Using Assay Employing Lipid-Coated Polystyrene Beads ‘f* Dorothy
Teegarden,
* Stephen
*Department of Foods and Nutrition, and fCommittee on Human Nutrition
Received
April
C. Meredith,?
and Michael
Purdue University, West Lafayette, and Nutritional Biology, University
0003-2697/92
Copyright All rights
IN 47907; and t Department of Pathology of Chicago, Chicago IL 60637
11, 1991
We have developed an assay to measure the affinity of serum vitamin D binding protein for 25-hydroxyvitamin DB, 1,25-dihydroxyvitamin D, , and vitamin D, , using uniform diameter (6.4 pm) polystyrene beads coated with phosphatidylcholine and vitamin D metabolites as the vitamin D donor. The lipid metabolite coated beads have a solid core, and thus all of the vitamin D metabolites are on the bead surface from which transfer to protein occurs. After incubating these beads in neutral buffer for 3 h, essentially no 3H-labeled vitamin D metabolites desorb from this surface. Phosphatidylcholine/vitamin D metabolite-coated beads ( 1 I.IM vitamin D metabolite) were incubated with varying concentrations of serum vitamin D binding protein under conditions in which the bead surfaces were saturated with protein, but most of the protein was free in solution. After incubation, beads were rapidly centrifuged without disturbing the equilibrium of binding and vitamin D metabolite bound to sDBP in solution was assayed in the supernatant. All three vitamin D metabolites became bound to serum vitamin D binding protein, and after 10 min of incubation the transfer of the metabolites to serum vitamin D binding protein was time independent. The transfer followed a Langmuir isotherm, and the Kd for each metabolite binding to serum vitamin D binding protein was derived by nonlinear leastsquares fit analysis. From this analysis the following values for the Kd were obtained: 5.59 X lo-’ M, 25-hydroxyvitamin D; 9.45 X 10m6 M, 1,25-dihydroxyvitamin D; and 9.17 X lo-’ M, vitamin D. In conclusion, we have developed a method which avoids problems encountered in previous assays and allows the precise and convenient determination of binding affinities of vita-
i This research Clinical Nutrition
D. Sitrin$
was supported by N.I.H. Research Unit DIDDK
$3.00
0 1992 by Academic Press, Inc. of reproduction in any form reserved.
Grant 1 P30 DK 42086 DK26678.
and
min D metabolites tein. 0 1991 Academic
and Press,
serum
vitamin
D binding
pro-
Inc.
Serum vitamin D2 binding protein (sDBP) plays an important role in transporting vitamin D, (D) and its metabolites, 25-hydroxyvitamin D, (250HD), and 1,25dihydroxyvitamin D, ( 1,25 (OH),D). Only l-4% of the binding capacity of sDBP is occupied by 250HD in the plasma at normal plasma vitamin D concentrations ( 1,2). sDBP has been found associated with lymphocyte or monocyte membranes and in the cytosol of a variety of lymphocyte populations and thus sDBP may be involved in the process of cellular association and internalization of 250HD (3-5). In addition to its role in vitamin D steroid transport, sDBP may also sequester free actin in the plasma (6-9). Further, sDBP can bind ligands other than D and D metabolites, such as unsaturated fatty acids, and thus may also function as a carrier for circulating fatty acids ( 10). Previous assays of the binding of D and D metabolites to sDBP have all been problematical, mainly in two areas: the method of introducing the ligand into the assay mixture, and the separation of bound and free vitamin after the incubation (1,2,11-18). Lipoproteins, detergents, and ethanol have been used by some investigators to solubilize D compounds, but these systems are very complex because the locus of the D metabolites may change within the lipoprotein as the concen’ Abbreviations used: D, vitamin D; 250HD, 25-hydroxyvitamin D; 1,25 (OH),D, 1,25dihydroxyvitamin D; PC, egg phosphatidylcholine or lecithin; sDBP, serum vitamin D binding protein; HPLC, high-performance liquid chromatography; TLC, thin-layer chromatography; SDS-PAGE, sodium dodecyl sulfate-Polyacrylamide gel electrophoresis. 293
294
TEEGARDEN,
MEREDITH,
tration of the ligand changes (11,14-16). In addition, detergents and ethanol can alter the structure of the protein being studied. One approach to separating bound and free D metabolites is to remove free vitamins by nonspecific adsorption to dextran-coated charcoal (1,X,16). This system is kinetically controlled, however, and never reaches equilibrium; 250HD is progressively lost from solution and increasing amounts of 250HD continuously adsorb onto the charcoal over time (13). Several other methods separate bound and free ligand using a sucrose gradient, a hydroxylapatite column, a DE81 filter, or centrifugal dialysis (12,14,16,18). As with the charcoal method, continuous nonspecific adsorption can continue over time and no true equilibrium is reached. In addition, some of these techniques are time consuming and would allow for repartitioning of ligand after the incubation period. We sought to develop an assay to facilitate the study of D and D metabolite binding to vitamin D binding proteins. The two main problems which need to be addressed in developing such an assay are that D compounds are only sparingly water soluble and they adsorb nonspecifically to a wide variety of surfaces. In this paper, we describe an assay which circumvents the problems associated with previous assays. D or D metabolites are incorporated into a mixed monolayer adsorbed to a solid support, uniform diameter (6.4 pm) polystyrene divinylbenzene beads. Thus, as we show, this assay system obviates the need to introduce D or D metabolites in soluble form with a carrier such as ethanol or detergents. In addition, separation of proteinbound and non-protein-bound fractions can be accomplished by a simple rapid centrifugation step which does not allow for repartitioning of the ligand.
METHODS
Materials Tris base, sodium dodecyl sulfate, glycine (electrophoand N, N, N’, N’-tetramethylethyleneretie grade), diamine (electrophoretic grade) were from Bethesda Research Laboratories (Gaithersburg, MD). 3- [NMorpholino] propanesulfonic acid (Mops) was received from Sigma. Acrylamide and N, N’-methylenebis acrylamide (electrophorectrophoretic grade) were from National Diagnostics (Somerville, NJ). Coomassie blue was from Bio-Rad (19). Other chemicals were of at least reagent grade. DEAE Sephacel was from Pharmacia Fine Chemicals. Agarose-hexylamine was from P-L Biochemical.% Inc. (Milwaukee, WI). The high-performance liquid chromatography (HPLC ) G3000SW gel permeation column was from ToyoSoda ( Bio-Rad) . Thin-layer chromatography (TLC) plates (LHP-KF linear K high-performance silica gel) were from Whatman. Polystyrene
AND
SITRIN
divinylbenzene beads (6.4 pm) were from Dow Diagnostic. Radioactive compounds were from Amersham. Unlabeled vitamin D, and 1,25 (OH),D, were from CalBiothem. 250HD, was a gift from J. A. Campbell of Upjohn ( Kalamazoo, MI ) . Male Sprague-Dawley rats weighing 150-400 g, used for purification of the serum proteins, were from Sprague-Dawley. Egg lecithin (PC) was from Avanti Polar Lipids. Radioactivity was measured in a Packard TriCarb scintillation counter. An IBM 9533 ternary gradient system was used for all HPLC analyses with a PerkinElmer LC-85 detector.
PC and Vitamin Method
D Metabolite
Polystyrene
Beads Assay
Methods for preparing colloidal silica-coated polystyrene divinylbenzene beads are previously described (20). Briefly, polystyrene divinylbenzene beads were cleaned with 20% sodium hydroxide (w/v) to remove the silica coating. The sodium hydroxide was removed by repeated washing with 10% ethanol until the pH was 7.0, and the beads were lyophilyzed. The D metabolites were purified before use with a SepPak column (21). Initially beads were prepared with 9 nmol of 250HD, a tracer dose of [ 3H] 250HD (specific activity = 5 X 10e5 nmol /DPM) , and 45 nmol PC in 100 ~1 ethanol to coat 50 mg of 6.4~pm beads in 0.5 ml hexane:ethanol, essentially as described (20). The beads were sonicated in a bath sonicator at 0.3 W /cm2 and dispersed with a pipet. The beads were then dried with nitrogen. Four milliliters of 0.02 M Mops, pH 7.0,0.16 M KC1 (Buffer A) was added and sonicated for 2 min while dispersing the beads with a pipet, and the beads were then centrifuged at 300g for 1 min. This wash procedure was repeated five times to remove excess PC. The number of beads per milliliter was calculated from the absorbance (A,,) of 50 ~1 of beads measured in a 3-ml solution of 2 mg/ml bovine serum albumin as previously described (20). The radioactivity of a 50-~1 aliquot was also measured to calculate the spreading density of the D metabolite on the surface of the beads from the specific radioactivity. In some experiments 14C-labeled PC with and without 250HD was used to coat the beads to measure the spreading densities of the PC. The integrity of each of the D metabolites and PC after these manipulations was assessed. An aliquot of beads was extracted several times with ethanol and washed, and the extracted material was analyzed by HPLC using a reverse-phase Cl8 column eluted with water:methanol at various ratios depending on the particular D metabolite being analyzed (Table 2). The purity of the PC was assessed by TLC as follows: Whatman LHP-KF linear K high-performance silica gel
BINDING
ASSAY
FOR
VITAMIN
plates of 200 pm thickness, approximately 3.5 X 10 cm, were spotted with 10 pg of PC or an extract. The plate was developed with chloroform:methanol:water (65:25:4) and dried, and the PC was visualized with iodine (22). The assay was done in 6 X 50-mm Kimble test tubes set in microfuge tubes as carriers in an Eppendorf microfuge (~5414). An aliquot of beads in buffer with the appropriate D or D metabolite concentration was incubated at a final volume of 0.4 ml at room temperature. After the incubation the samples were centrifuged for 10 s at 17,000g. Two 100~~1 aliquots of the supernatant were sampled and the tube was rinsed with fluor to measure the radioactivity left in the tube. Purified sDBP was added in concentrations of 0.13 to 10.00 pM depending on the experiment. Purification
of Serum
Vitamin
D Binding
Protein
The sDBP was purified from male Sprague-Dawley rats. The protein was purified essentially as previously described (23), with minor modifications, by ammonium sulfate precipitation, gel filtration chromatography, DEAE Sephacel ion exchange chromatography, and a 250HD affinity column. The 40-60% ammonium sulfate saturated precipitate was resolubolized in 0.1 M sodium chloride, 0.05 M sodium phosphate, pH 6.8. A tracer amount of [ 3H] 250HD was added to follow the elution of the binding protein in the following two chromatographic steps. The ammonium sulfate fraction was applied to a ToyoSoda G3000SW gel permeation column in an HPLC system. The column was eluted with the same buffer at 1 ml /min. The sDBP was detected in a single peak of radioactivity of an apparent molecular weight of 58,000. sDBP obtained from the gel permeation column was dialyzed against 0.1 M NaCl, 0.01 M Tris, pH 7.4, in preparation for the DEAE Sephecal anion exchange column. The sample was applied to the column and eluted using a O-l M sodium chloride gradient; the sDBP eluted at 0.2 M sodium chloride. The sDBP obtained from the previous chromatographic procedures was further purified using a 250HD affinity column prepared according to the method of Haddad ( 23). The purity as assessed by SDS-PAGE with silver staining of a gel was >98% (24). Purified sDBP was solubilized in water, divided into small fractions, and stored at -80°C. The protein concentration of purified sDBP aliquots and during experiments was determined by the Bio-Rad method (19). RESULTS
Characterization of Assay Uniform diameter polystyrene divinylbenzene beads were used by Retzinger, et al. (20) to assessthe binding of amphiphilic peptides to lipid surfaces. In this paper,
D AND
BINDING
PROTEINS
295
the rationale of using these beads coated with PC and D metabolites was that D metabolites would adsorb to, and remain bound to, the surface of phosphatidylcholine-coated polystyrene divinylbenzene beads. The beads then could be used as a reservoir for D metabolites in assays to measure the binding of D metabolites to sDBP, and since there is no ligand hydrophobic phase, all interactions would take place either at the surface or in the solution phase. The beads were prepared with a mixed monolayer of PC and D metabolites according to the method of Retzinger for coating with PC alone. The D metabolite /PC molar ratio in ethanol used to prepared the mixed monolayer was 0.2. This ratio resembles somewhat the cholesterol/PC ratio in lipoproteins. To 50-mg beads in hexane:ethanol (v iv), 9 nmol of D metabolite and 45 nmol of PC in ethanol were added. After sonication and drying with nitrogen, the beads were sonicated and washed four times with Buffer A to remove excess PC. Previous studies of PC-coated polystyrene divinylbenzene beads have shown that sonication and washing removed any PC above an equilibrium spreading density of 150 A” /molecule. Our current studies (below ) demonstrate that the addition of D or D metabolite to PC does not alter the final surface density of the PC molecule. After four washes, the PC remained adsorbed on the surface of the beads and small amounts of metabolite were lost in the washes but reached final molar ratios (D /PC) of approximately 0.13, 0.12, and 0.09 for D, 250HD, and 1,25 (OH),D, respectively. Thus, 30 nmol of the PC was recovered on the surface of the beads when 45 nmol was used to prepare 50 mg of beads, and approximately 3.9, 3.6, and 2.7 nmol of D, 250HD, and 1,25 ( OH )2D were recovered on the surface of the beads when 9 nmol of metabolite was used to prepare 50 mg of beads. In another experiment, the effect of varying the initial 250HD /PC ratio was examined by altering the level of PC used such that the ratios were 0.2,0.15, and 0.075. In each case, varying the initial 250HD /PC ratio did not significantly alter the final 250HD /PC ratio recovered on the beads of 0.12. The equilibrium spreading density of D metabolites D, 250HD, and 1,25(OH),D on the bead surface was calculated from the surface area of the beads, the radioactivity of the D and PC, and the specific radioactivity of the compounds; these data are shown in Table 1. The equilibrium spreading density of PC on the surface of the beads was also determined with and without 250HD also incorporated on the surface of the beads (Table 1) . These results confirm that 250HD adsorbs without altering the surface concentration of the PC molecules on the surface of the beads, suggesting that it adsorbs onto the hydrophobic domains between the PC polar head groups. In subsequent experiments, the nominal concentration of D metabolite will depend on the moles of D metabolite in the suspension, divided by the total vol-
296
TEEGARDEN, TABLE
Spreading
Density Metabolites
MEREDITH,
AND
SITRIN
1
of Phosphatidylcholine on Polystyrene Beads
and D
Spreading Density @*/molecule)
Metabolite Vitamin D 250HD 1,25(OH),D PC (with 250HD) PC (without 250HD)
% 2
909 1028 1392 115 112
n B 2
ume of the suspension (L) , since the volume percent of the suspension occupied by the beads is negligible ( ~0.1% ) . The nominal concentration chosen for the assay was 1 PM metabolite. Total surface area of beads used in the assay was approximately 32 cm2. The PC/D metabolite-coated beads were mixed with Buffer A to a final volume of 400 ~1 with a nominal concentration of 1 pM D or D metabolite. The concentration of D or D metabolite was measured in the slurry of beads and in the supernatant after centrifugation as a function of time. The data in Fig. 1 show that the D metabolites do not desorb from the surface over time in the absence of sDBP. The PC and D metabolite-coated beads were extracted and the integrity of the D metabolites was assessedby analytical reverse-phase HPLC. The integrity was assessedby calculating the percentage recovery of the metabolite label coeluting with a standard on a reverse-phase HPLC column (Table 2). In each case the integrity of the D metabolites was assessedto be greater than 92%. Following extraction, the integrity of the PC on beads coated with only PC (i.e., without D) was assessed by TLC, and a single spot coeluting with PC standard was visualized by exposing the plate to iodine vapor. Since amphiphilic peptides adsorb onto the surface of the beads, the binding of sDBP to the PC- and 250HDcoated beads during the experiments was measured. Beads coated with 250HD /PC at a nominal 250HD concentration of 1 pM and having a total bead surface area of 32 cm2 were incubated for 1 h at 25°C with 25 to 100 /*g sDBP. In a parallel experiment, 32 cm2 of beads coated with only PC was incubated with the sDBP. The
0.20 0.00 1.00 0.80
B
---“-“.-*.--‘....-*---.------.
0.60 0.40 0.20 0.00
30
0
60
TIME
FIG. 1. Desorption of D metabolites from the surface of the beads. Coated PC/D metabolite beads were incubated in the absence of protein, and aliquots of the bead slurry (dashed lines) and the supernatant (solid lines) after centrifugation were sampled. (A) (B) and (C) are D, 250HD and 1,25( OH),D, respectively.
amount of protein in the supernatant at the end of the incubation was measured using the Bio-Rad method (19). Approximately 3 pg of sDBP bound to the PCcoated beads, with or without 250HD on the surface, at all concentrations of sDBP, indicating that the surface of the beads was saturated by a small amount of sDBP. Minimal protein was lost from solution, i.e., adsorbed to the bead surface, during the assay with the protein concentrations generally used. The results described above demonstrate that the PC /D metabolite-coated bead assay is appropriate for studying D metabolite binding to sDBP: the metabolites adsorb to the surface and then do not desorb from the surface of the beads over time in the absence of sDBP; the D metabolites and PC retain their integrity, and the surface of the PC/D metabolite-coated beads becomes
TABLE
Parameters
Determined
3
for Binding
Integrity Metabolite D 250HD 1,25(OH),D
2
of D and D Metabolites %Recovery coeluting
of radiolabel with standards 92 95 93
& (PM) SD of Kd
after Incubation Eluting MeOH:H,O
D tot.; (PM) solvent (v/v)
96:4 85:15 8515
Isotherms
Metabolite D
TABLE
90
(minutes)
SD of” AICb
ss a L, = Dtowmaai,~e. ’ Akaike Information ‘Sum of Squares.
250HD
9.17 4.18
5.59 0.88
0.202 0.057 -68.66 7 x lo-'
0.390 0.038 -78.92 3 x lo-’
criterion
(25).
1,25(OH),D 9.45
4.77 0.636 0.208 -64.02 9 x lo-’
BINDING
ASSAY
FOR
VITAMIN
D AND
BINDING
297
PROTEINS
available in the assay mix will be called D,, a”. The concentration (M) of D metabolite which is not bound to protein, but which remains available for binding to protein (D,, ..) will then be: D i-eB” = Dtotav - Cx. 0.04 P 3 0 2
Thus, the amount of D metabolites solution can be expressed by:
0.03 0.02
gz
0.01 1 n
a
01,
,, 0
,,
,,,,,l,
10
20
),, 30 40 TIME (minutes)
,,,
, I, 50
60
,I,,
70
/.I, 80
saturated with a minimal amount of sDBP. The assay was done with a final volume of 400 ~1 and a nominal concentration of 1 PM D metabolite. After incubation, the supernatant was sampled to measure the radioactivity, and the vitamin D concentration in the supernatant was determined using the specific activity of the radiolabeled metabolite and assuming that the beads occupied minimal volume.
where Sfree = concentration ( M) of vacant D metabolite binding sites on sDBP and Kd = dissociation constant (M).LetS,,,,= total concentration (M) of D metabolite binding sites on sDBP, whether free or occupied. Since s free = &.0t*1 - Cx and Stota, P Cx under the experimental conditions used, we can make the approximation: S free Thus, the expression
to Vitamin
=
StotA.
for Kd,
Kd = (D,, av) ( Site,, can be put into experimentally &
and 1,250HD
in
I 90
FIG. 2. Binding of D metabolites to sDBP as a function of time. A nominal concentration of 1 pM D or D metabolite on PC/D metabolite-coated beads were incubated with 1 pM sDBP from 2.5 to 90 min at 25°C. The concentration of bound metabolites, D (open square), 250HD (closed circle), and 1,25(OH&D (open circle), in the supernatant was determined.
Binding of D, 25OHD, Binding Protein
bound to sDBP
) 1 Cx ,
obtainable
terms:
= (Dtotav - Cx)(S,,,)/Cx.
D
When a nominal D metabolite concentration of 1 PM PC /D metabolite-coated beads was incubated with 1 PM sDBP, the D metabolite associated with the sDBP increased and reached a plateau within 10 min (Fig. 2). The kinetics of this transfer are consistent with a monoexponential approach to equilibrium. For future experiments, a l-h incubation period was selected. We used the PC /D metabolite coated beads to determine the binding affinity of sDBP for D and D metabolites. PC/D metabolite-coated beads were suspended in Buffer A at 25°C to yield a nominal concentration of D metabolite of 1 PM. Various concentrations of sDBP, ranging from 0.143 to 5.0 /IM were then added to the bead suspension, and the mixture was incubated for 1 h at 25°C. After the incubation, the beads were centrifuged at 17,000g for 10 s and the concentration of D metabolite in the supernatant was determined. If D,, = nominal concentration of D metabolite in the bead suspension, then Dbw aveilable= xD,~, where 0 < x < 1 and x = fraction of Dtot potentially available for binding to sDBP. At a given concentration of sDBP, the concentration (M) of D metabolite bound to sDBP in solution will be called Cx = complex, of D and sDBP, and the total D
Several independent investigations have indicated that sDBP makes a 1:l stoichiometric complex with D metabolites; thus, S,,, is simply the concentration (M) of sDBP in solution ( 11,12,26-28). The CX concentration is thus equal to the concentration of D metabolite in the supernatant after centrifugation of the bead suspension, since the concentration of D metabolite free in solution is negligible (
BIOCHEMISTRY
199,
293-299
( 1991)
Determination of the Affinity of Vitamin D Metabolites to Serum Vitamin D Binding Protein Using Assay Employing Lipid-Coated Polystyrene Beads ‘f* Dorothy
Teegarden,
* Stephen
*Department of Foods and Nutrition, and fCommittee on Human Nutrition
Received
April
C. Meredith,?
and Michael
Purdue University, West Lafayette, and Nutritional Biology, University
0003-2697/92
Copyright All rights
IN 47907; and t Department of Pathology of Chicago, Chicago IL 60637
11, 1991
We have developed an assay to measure the affinity of serum vitamin D binding protein for 25-hydroxyvitamin DB, 1,25-dihydroxyvitamin D, , and vitamin D, , using uniform diameter (6.4 pm) polystyrene beads coated with phosphatidylcholine and vitamin D metabolites as the vitamin D donor. The lipid metabolite coated beads have a solid core, and thus all of the vitamin D metabolites are on the bead surface from which transfer to protein occurs. After incubating these beads in neutral buffer for 3 h, essentially no 3H-labeled vitamin D metabolites desorb from this surface. Phosphatidylcholine/vitamin D metabolite-coated beads ( 1 I.IM vitamin D metabolite) were incubated with varying concentrations of serum vitamin D binding protein under conditions in which the bead surfaces were saturated with protein, but most of the protein was free in solution. After incubation, beads were rapidly centrifuged without disturbing the equilibrium of binding and vitamin D metabolite bound to sDBP in solution was assayed in the supernatant. All three vitamin D metabolites became bound to serum vitamin D binding protein, and after 10 min of incubation the transfer of the metabolites to serum vitamin D binding protein was time independent. The transfer followed a Langmuir isotherm, and the Kd for each metabolite binding to serum vitamin D binding protein was derived by nonlinear leastsquares fit analysis. From this analysis the following values for the Kd were obtained: 5.59 X lo-’ M, 25-hydroxyvitamin D; 9.45 X 10m6 M, 1,25-dihydroxyvitamin D; and 9.17 X lo-’ M, vitamin D. In conclusion, we have developed a method which avoids problems encountered in previous assays and allows the precise and convenient determination of binding affinities of vita-
i This research Clinical Nutrition
D. Sitrin$
was supported by N.I.H. Research Unit DIDDK
$3.00
0 1992 by Academic Press, Inc. of reproduction in any form reserved.
Grant 1 P30 DK 42086 DK26678.
and
min D metabolites tein. 0 1991 Academic
and Press,
serum
vitamin
D binding
pro-
Inc.
Serum vitamin D2 binding protein (sDBP) plays an important role in transporting vitamin D, (D) and its metabolites, 25-hydroxyvitamin D, (250HD), and 1,25dihydroxyvitamin D, ( 1,25 (OH),D). Only l-4% of the binding capacity of sDBP is occupied by 250HD in the plasma at normal plasma vitamin D concentrations ( 1,2). sDBP has been found associated with lymphocyte or monocyte membranes and in the cytosol of a variety of lymphocyte populations and thus sDBP may be involved in the process of cellular association and internalization of 250HD (3-5). In addition to its role in vitamin D steroid transport, sDBP may also sequester free actin in the plasma (6-9). Further, sDBP can bind ligands other than D and D metabolites, such as unsaturated fatty acids, and thus may also function as a carrier for circulating fatty acids ( 10). Previous assays of the binding of D and D metabolites to sDBP have all been problematical, mainly in two areas: the method of introducing the ligand into the assay mixture, and the separation of bound and free vitamin after the incubation (1,2,11-18). Lipoproteins, detergents, and ethanol have been used by some investigators to solubilize D compounds, but these systems are very complex because the locus of the D metabolites may change within the lipoprotein as the concen’ Abbreviations used: D, vitamin D; 250HD, 25-hydroxyvitamin D; 1,25 (OH),D, 1,25dihydroxyvitamin D; PC, egg phosphatidylcholine or lecithin; sDBP, serum vitamin D binding protein; HPLC, high-performance liquid chromatography; TLC, thin-layer chromatography; SDS-PAGE, sodium dodecyl sulfate-Polyacrylamide gel electrophoresis. 293
294
TEEGARDEN,
MEREDITH,
tration of the ligand changes (11,14-16). In addition, detergents and ethanol can alter the structure of the protein being studied. One approach to separating bound and free D metabolites is to remove free vitamins by nonspecific adsorption to dextran-coated charcoal (1,X,16). This system is kinetically controlled, however, and never reaches equilibrium; 250HD is progressively lost from solution and increasing amounts of 250HD continuously adsorb onto the charcoal over time (13). Several other methods separate bound and free ligand using a sucrose gradient, a hydroxylapatite column, a DE81 filter, or centrifugal dialysis (12,14,16,18). As with the charcoal method, continuous nonspecific adsorption can continue over time and no true equilibrium is reached. In addition, some of these techniques are time consuming and would allow for repartitioning of ligand after the incubation period. We sought to develop an assay to facilitate the study of D and D metabolite binding to vitamin D binding proteins. The two main problems which need to be addressed in developing such an assay are that D compounds are only sparingly water soluble and they adsorb nonspecifically to a wide variety of surfaces. In this paper, we describe an assay which circumvents the problems associated with previous assays. D or D metabolites are incorporated into a mixed monolayer adsorbed to a solid support, uniform diameter (6.4 pm) polystyrene divinylbenzene beads. Thus, as we show, this assay system obviates the need to introduce D or D metabolites in soluble form with a carrier such as ethanol or detergents. In addition, separation of proteinbound and non-protein-bound fractions can be accomplished by a simple rapid centrifugation step which does not allow for repartitioning of the ligand.
METHODS
Materials Tris base, sodium dodecyl sulfate, glycine (electrophoand N, N, N’, N’-tetramethylethyleneretie grade), diamine (electrophoretic grade) were from Bethesda Research Laboratories (Gaithersburg, MD). 3- [NMorpholino] propanesulfonic acid (Mops) was received from Sigma. Acrylamide and N, N’-methylenebis acrylamide (electrophorectrophoretic grade) were from National Diagnostics (Somerville, NJ). Coomassie blue was from Bio-Rad (19). Other chemicals were of at least reagent grade. DEAE Sephacel was from Pharmacia Fine Chemicals. Agarose-hexylamine was from P-L Biochemical.% Inc. (Milwaukee, WI). The high-performance liquid chromatography (HPLC ) G3000SW gel permeation column was from ToyoSoda ( Bio-Rad) . Thin-layer chromatography (TLC) plates (LHP-KF linear K high-performance silica gel) were from Whatman. Polystyrene
AND
SITRIN
divinylbenzene beads (6.4 pm) were from Dow Diagnostic. Radioactive compounds were from Amersham. Unlabeled vitamin D, and 1,25 (OH),D, were from CalBiothem. 250HD, was a gift from J. A. Campbell of Upjohn ( Kalamazoo, MI ) . Male Sprague-Dawley rats weighing 150-400 g, used for purification of the serum proteins, were from Sprague-Dawley. Egg lecithin (PC) was from Avanti Polar Lipids. Radioactivity was measured in a Packard TriCarb scintillation counter. An IBM 9533 ternary gradient system was used for all HPLC analyses with a PerkinElmer LC-85 detector.
PC and Vitamin Method
D Metabolite
Polystyrene
Beads Assay
Methods for preparing colloidal silica-coated polystyrene divinylbenzene beads are previously described (20). Briefly, polystyrene divinylbenzene beads were cleaned with 20% sodium hydroxide (w/v) to remove the silica coating. The sodium hydroxide was removed by repeated washing with 10% ethanol until the pH was 7.0, and the beads were lyophilyzed. The D metabolites were purified before use with a SepPak column (21). Initially beads were prepared with 9 nmol of 250HD, a tracer dose of [ 3H] 250HD (specific activity = 5 X 10e5 nmol /DPM) , and 45 nmol PC in 100 ~1 ethanol to coat 50 mg of 6.4~pm beads in 0.5 ml hexane:ethanol, essentially as described (20). The beads were sonicated in a bath sonicator at 0.3 W /cm2 and dispersed with a pipet. The beads were then dried with nitrogen. Four milliliters of 0.02 M Mops, pH 7.0,0.16 M KC1 (Buffer A) was added and sonicated for 2 min while dispersing the beads with a pipet, and the beads were then centrifuged at 300g for 1 min. This wash procedure was repeated five times to remove excess PC. The number of beads per milliliter was calculated from the absorbance (A,,) of 50 ~1 of beads measured in a 3-ml solution of 2 mg/ml bovine serum albumin as previously described (20). The radioactivity of a 50-~1 aliquot was also measured to calculate the spreading density of the D metabolite on the surface of the beads from the specific radioactivity. In some experiments 14C-labeled PC with and without 250HD was used to coat the beads to measure the spreading densities of the PC. The integrity of each of the D metabolites and PC after these manipulations was assessed. An aliquot of beads was extracted several times with ethanol and washed, and the extracted material was analyzed by HPLC using a reverse-phase Cl8 column eluted with water:methanol at various ratios depending on the particular D metabolite being analyzed (Table 2). The purity of the PC was assessed by TLC as follows: Whatman LHP-KF linear K high-performance silica gel
BINDING
ASSAY
FOR
VITAMIN
plates of 200 pm thickness, approximately 3.5 X 10 cm, were spotted with 10 pg of PC or an extract. The plate was developed with chloroform:methanol:water (65:25:4) and dried, and the PC was visualized with iodine (22). The assay was done in 6 X 50-mm Kimble test tubes set in microfuge tubes as carriers in an Eppendorf microfuge (~5414). An aliquot of beads in buffer with the appropriate D or D metabolite concentration was incubated at a final volume of 0.4 ml at room temperature. After the incubation the samples were centrifuged for 10 s at 17,000g. Two 100~~1 aliquots of the supernatant were sampled and the tube was rinsed with fluor to measure the radioactivity left in the tube. Purified sDBP was added in concentrations of 0.13 to 10.00 pM depending on the experiment. Purification
of Serum
Vitamin
D Binding
Protein
The sDBP was purified from male Sprague-Dawley rats. The protein was purified essentially as previously described (23), with minor modifications, by ammonium sulfate precipitation, gel filtration chromatography, DEAE Sephacel ion exchange chromatography, and a 250HD affinity column. The 40-60% ammonium sulfate saturated precipitate was resolubolized in 0.1 M sodium chloride, 0.05 M sodium phosphate, pH 6.8. A tracer amount of [ 3H] 250HD was added to follow the elution of the binding protein in the following two chromatographic steps. The ammonium sulfate fraction was applied to a ToyoSoda G3000SW gel permeation column in an HPLC system. The column was eluted with the same buffer at 1 ml /min. The sDBP was detected in a single peak of radioactivity of an apparent molecular weight of 58,000. sDBP obtained from the gel permeation column was dialyzed against 0.1 M NaCl, 0.01 M Tris, pH 7.4, in preparation for the DEAE Sephecal anion exchange column. The sample was applied to the column and eluted using a O-l M sodium chloride gradient; the sDBP eluted at 0.2 M sodium chloride. The sDBP obtained from the previous chromatographic procedures was further purified using a 250HD affinity column prepared according to the method of Haddad ( 23). The purity as assessed by SDS-PAGE with silver staining of a gel was >98% (24). Purified sDBP was solubilized in water, divided into small fractions, and stored at -80°C. The protein concentration of purified sDBP aliquots and during experiments was determined by the Bio-Rad method (19). RESULTS
Characterization of Assay Uniform diameter polystyrene divinylbenzene beads were used by Retzinger, et al. (20) to assessthe binding of amphiphilic peptides to lipid surfaces. In this paper,
D AND
BINDING
PROTEINS
295
the rationale of using these beads coated with PC and D metabolites was that D metabolites would adsorb to, and remain bound to, the surface of phosphatidylcholine-coated polystyrene divinylbenzene beads. The beads then could be used as a reservoir for D metabolites in assays to measure the binding of D metabolites to sDBP, and since there is no ligand hydrophobic phase, all interactions would take place either at the surface or in the solution phase. The beads were prepared with a mixed monolayer of PC and D metabolites according to the method of Retzinger for coating with PC alone. The D metabolite /PC molar ratio in ethanol used to prepared the mixed monolayer was 0.2. This ratio resembles somewhat the cholesterol/PC ratio in lipoproteins. To 50-mg beads in hexane:ethanol (v iv), 9 nmol of D metabolite and 45 nmol of PC in ethanol were added. After sonication and drying with nitrogen, the beads were sonicated and washed four times with Buffer A to remove excess PC. Previous studies of PC-coated polystyrene divinylbenzene beads have shown that sonication and washing removed any PC above an equilibrium spreading density of 150 A” /molecule. Our current studies (below ) demonstrate that the addition of D or D metabolite to PC does not alter the final surface density of the PC molecule. After four washes, the PC remained adsorbed on the surface of the beads and small amounts of metabolite were lost in the washes but reached final molar ratios (D /PC) of approximately 0.13, 0.12, and 0.09 for D, 250HD, and 1,25 (OH),D, respectively. Thus, 30 nmol of the PC was recovered on the surface of the beads when 45 nmol was used to prepare 50 mg of beads, and approximately 3.9, 3.6, and 2.7 nmol of D, 250HD, and 1,25 ( OH )2D were recovered on the surface of the beads when 9 nmol of metabolite was used to prepare 50 mg of beads. In another experiment, the effect of varying the initial 250HD /PC ratio was examined by altering the level of PC used such that the ratios were 0.2,0.15, and 0.075. In each case, varying the initial 250HD /PC ratio did not significantly alter the final 250HD /PC ratio recovered on the beads of 0.12. The equilibrium spreading density of D metabolites D, 250HD, and 1,25(OH),D on the bead surface was calculated from the surface area of the beads, the radioactivity of the D and PC, and the specific radioactivity of the compounds; these data are shown in Table 1. The equilibrium spreading density of PC on the surface of the beads was also determined with and without 250HD also incorporated on the surface of the beads (Table 1) . These results confirm that 250HD adsorbs without altering the surface concentration of the PC molecules on the surface of the beads, suggesting that it adsorbs onto the hydrophobic domains between the PC polar head groups. In subsequent experiments, the nominal concentration of D metabolite will depend on the moles of D metabolite in the suspension, divided by the total vol-
296
TEEGARDEN, TABLE
Spreading
Density Metabolites
MEREDITH,
AND
SITRIN
1
of Phosphatidylcholine on Polystyrene Beads
and D
Spreading Density @*/molecule)
Metabolite Vitamin D 250HD 1,25(OH),D PC (with 250HD) PC (without 250HD)
% 2
909 1028 1392 115 112
n B 2
ume of the suspension (L) , since the volume percent of the suspension occupied by the beads is negligible ( ~0.1% ) . The nominal concentration chosen for the assay was 1 PM metabolite. Total surface area of beads used in the assay was approximately 32 cm2. The PC/D metabolite-coated beads were mixed with Buffer A to a final volume of 400 ~1 with a nominal concentration of 1 pM D or D metabolite. The concentration of D or D metabolite was measured in the slurry of beads and in the supernatant after centrifugation as a function of time. The data in Fig. 1 show that the D metabolites do not desorb from the surface over time in the absence of sDBP. The PC and D metabolite-coated beads were extracted and the integrity of the D metabolites was assessedby analytical reverse-phase HPLC. The integrity was assessedby calculating the percentage recovery of the metabolite label coeluting with a standard on a reverse-phase HPLC column (Table 2). In each case the integrity of the D metabolites was assessedto be greater than 92%. Following extraction, the integrity of the PC on beads coated with only PC (i.e., without D) was assessed by TLC, and a single spot coeluting with PC standard was visualized by exposing the plate to iodine vapor. Since amphiphilic peptides adsorb onto the surface of the beads, the binding of sDBP to the PC- and 250HDcoated beads during the experiments was measured. Beads coated with 250HD /PC at a nominal 250HD concentration of 1 pM and having a total bead surface area of 32 cm2 were incubated for 1 h at 25°C with 25 to 100 /*g sDBP. In a parallel experiment, 32 cm2 of beads coated with only PC was incubated with the sDBP. The
0.20 0.00 1.00 0.80
B
---“-“.-*.--‘....-*---.------.
0.60 0.40 0.20 0.00
30
0
60
TIME
FIG. 1. Desorption of D metabolites from the surface of the beads. Coated PC/D metabolite beads were incubated in the absence of protein, and aliquots of the bead slurry (dashed lines) and the supernatant (solid lines) after centrifugation were sampled. (A) (B) and (C) are D, 250HD and 1,25( OH),D, respectively.
amount of protein in the supernatant at the end of the incubation was measured using the Bio-Rad method (19). Approximately 3 pg of sDBP bound to the PCcoated beads, with or without 250HD on the surface, at all concentrations of sDBP, indicating that the surface of the beads was saturated by a small amount of sDBP. Minimal protein was lost from solution, i.e., adsorbed to the bead surface, during the assay with the protein concentrations generally used. The results described above demonstrate that the PC /D metabolite-coated bead assay is appropriate for studying D metabolite binding to sDBP: the metabolites adsorb to the surface and then do not desorb from the surface of the beads over time in the absence of sDBP; the D metabolites and PC retain their integrity, and the surface of the PC/D metabolite-coated beads becomes
TABLE
Parameters
Determined
3
for Binding
Integrity Metabolite D 250HD 1,25(OH),D
2
of D and D Metabolites %Recovery coeluting
of radiolabel with standards 92 95 93
& (PM) SD of Kd
after Incubation Eluting MeOH:H,O
D tot.; (PM) solvent (v/v)
96:4 85:15 8515
Isotherms
Metabolite D
TABLE
90
(minutes)
SD of” AICb
ss a L, = Dtowmaai,~e. ’ Akaike Information ‘Sum of Squares.
250HD
9.17 4.18
5.59 0.88
0.202 0.057 -68.66 7 x lo-'
0.390 0.038 -78.92 3 x lo-’
criterion
(25).
1,25(OH),D 9.45
4.77 0.636 0.208 -64.02 9 x lo-’
BINDING
ASSAY
FOR
VITAMIN
D AND
BINDING
297
PROTEINS
available in the assay mix will be called D,, a”. The concentration (M) of D metabolite which is not bound to protein, but which remains available for binding to protein (D,, ..) will then be: D i-eB” = Dtotav - Cx. 0.04 P 3 0 2
Thus, the amount of D metabolites solution can be expressed by:
0.03 0.02
gz
0.01 1 n
a
01,
,, 0
,,
,,,,,l,
10
20
),, 30 40 TIME (minutes)
,,,
, I, 50
60
,I,,
70
/.I, 80
saturated with a minimal amount of sDBP. The assay was done with a final volume of 400 ~1 and a nominal concentration of 1 PM D metabolite. After incubation, the supernatant was sampled to measure the radioactivity, and the vitamin D concentration in the supernatant was determined using the specific activity of the radiolabeled metabolite and assuming that the beads occupied minimal volume.
where Sfree = concentration ( M) of vacant D metabolite binding sites on sDBP and Kd = dissociation constant (M).LetS,,,,= total concentration (M) of D metabolite binding sites on sDBP, whether free or occupied. Since s free = &.0t*1 - Cx and Stota, P Cx under the experimental conditions used, we can make the approximation: S free Thus, the expression
to Vitamin
=
StotA.
for Kd,
Kd = (D,, av) ( Site,, can be put into experimentally &
and 1,250HD
in
I 90
FIG. 2. Binding of D metabolites to sDBP as a function of time. A nominal concentration of 1 pM D or D metabolite on PC/D metabolite-coated beads were incubated with 1 pM sDBP from 2.5 to 90 min at 25°C. The concentration of bound metabolites, D (open square), 250HD (closed circle), and 1,25(OH&D (open circle), in the supernatant was determined.
Binding of D, 25OHD, Binding Protein
bound to sDBP
) 1 Cx ,
obtainable
terms:
= (Dtotav - Cx)(S,,,)/Cx.
D
When a nominal D metabolite concentration of 1 PM PC /D metabolite-coated beads was incubated with 1 PM sDBP, the D metabolite associated with the sDBP increased and reached a plateau within 10 min (Fig. 2). The kinetics of this transfer are consistent with a monoexponential approach to equilibrium. For future experiments, a l-h incubation period was selected. We used the PC /D metabolite coated beads to determine the binding affinity of sDBP for D and D metabolites. PC/D metabolite-coated beads were suspended in Buffer A at 25°C to yield a nominal concentration of D metabolite of 1 PM. Various concentrations of sDBP, ranging from 0.143 to 5.0 /IM were then added to the bead suspension, and the mixture was incubated for 1 h at 25°C. After the incubation, the beads were centrifuged at 17,000g for 10 s and the concentration of D metabolite in the supernatant was determined. If D,, = nominal concentration of D metabolite in the bead suspension, then Dbw aveilable= xD,~, where 0 < x < 1 and x = fraction of Dtot potentially available for binding to sDBP. At a given concentration of sDBP, the concentration (M) of D metabolite bound to sDBP in solution will be called Cx = complex, of D and sDBP, and the total D
Several independent investigations have indicated that sDBP makes a 1:l stoichiometric complex with D metabolites; thus, S,,, is simply the concentration (M) of sDBP in solution ( 11,12,26-28). The CX concentration is thus equal to the concentration of D metabolite in the supernatant after centrifugation of the bead suspension, since the concentration of D metabolite free in solution is negligible (
