- Email: [email protected]
Biochemical and preliminary crystallographic characterization of the vitamin D sterol- and actin-binding by human vitamin D-binding protein
Archives of Biochemistry and Biophysics 402 (2002) 14–23 www.academicpress.com
Biochemical and preliminary crystallographic characterization of the vitamin D sterol- and actin-binding by human vitamin D-binding protein Narasimha Swamy,a James F. Head,b Daniel Weitz,a and Rahul Raya,* a
Bioorganic Chemistry & Structural Biology, Section in Endocrinology, Diabetes and Metabolism, Department of Medicine, Boston University School of Medicine, Boston, MA 02118, USA b Department of Physiology & Biophysics, Boston University School of Medicine, Boston, MA 02118, USA Received 20 December 2001, and in revised form 19 February 2002
Abstract Vitamin D-binding protein (DBP), a multi-functional serum glycoprotein, has a triple-domain modular structure. Mutation of Trp145 (in Domain I) to Ser decreased 25-OH-D3 -binding by 80%. Furthermore, recombinant Domain I (1–203) and Domain I + II (1–330) showed specific and strong binding for 25-OH-D3 , but Domain III (375–427) did not, suggesting that only Domains I and II might be required for vitamin D sterol-binding. Past studies have suggested that Domain III is independently capable of binding G-actin. We exploited this apparently independent ligand-binding property of DBP to purify DBP–actin complex from human serum and rabbit muscle actin by 25-OH-D3 affinity chromatography. Competitive 3 H-25-OH-D3 binding curves for native DBP and DBP–actin complex were almost identical, further suggesting that vitamin D sterol- and actin-binding activities by DBP might be largely independent of each other. Trypsin treatment of DBP produced a prominent 25 kDa band (Domain I, minus 5 amino acids in N-terminus), while actin was completely fragmented by such treatment. In contrast, tryptic digestion of purified DBP–actin complex showed two prominent bands, 52 (DBP, minus 5 amino acids in the N-terminus) and 34 kDa (actin, starting with amino acid position 69) indicating that DBP, particularly its Domains II and III were protected from trypsin cleavage upon actin-binding. Similarly, actin, except its N-terminus, was also protected from tryptic digestion when complexed with DBP. These results provided the basis crystal, primitive orthorhombic with unit cell dimensions for our studies to crystallize DBP–actin complex, which produced a 2.5 A , b ¼ 87:3 A , and c ¼ 159:6 A , P21 21 21 space group, Vm ¼ 2:9. Soaking of crystals of actin–DBP in crystallization buffer a ¼ 80:2 A containing various concentrations of 25-OH-D3 resulted in cracking of the crystal, which was probably a reflection of a ligandinduced conformational change in the complex, disrupting crystal contacts. In conclusion, we have provided data to suggest that although binding of 25-OH-D3 to DBP might result in discrete conformational changes in the holo-protein to influence actin-binding, these binding processes are largely independent of each other in solution. Ó 2002 Elsevier Science (USA). All rights reserved. Keywords: Vitamin D-binding protein; G-actin; 25-hydroxyvitamin D3 ; 1,25-dihydroxyvitamin D3 ; DBP-mutant; Recombinant DBP-structural domains; DBP–actin crystal
Vitamin D-binding protein (DBP)1 is a polymorphic serum glycoprotein with multiple functions that include highly specific binding of vitamin D sterols, G-actin, fatty acids, and chemotactic agents [1–3]. Furthermore, *
Corresponding author. Fax: +617-638-8194. E-mail address: [email protected] (R. Ray). 1 Abbreviations used: hDBP, human vitamin D-binding protein; DBP-maf, DBP-macrophage-activating factor; 25-OH-D3 , 25-hydroxyvitamin D3 ; 1; 25ðOHÞ2 D3 , la,25-dihydroxyvitamin D3 ; reDBP, recombinant DBP.
DBP undergoes a sequential removal of sugar moieties (by glycosidases produced by inflammation-primed B- and T-lymphocytes) to generate a posttranslationally modified form of DBP (DBP-macrophage-activating factor, DBP-maf), which is a potent activator of macrophages and osteoclasts [4–7]. Furthermore, DBP-maf was found to increase the survival rate of mice bearing Ehrlich ascites tumor [8,9]. Vitamin D sterol transport by DBP has been studied extensively [1–3]. DBP binds vitamin D3 in serum and transports this seco-steroid to liver to form 25-hydroxy-
0003-9861/02/$ - see front matter Ó 2002 Elsevier Science (USA). All rights reserved. PII: S 0 0 0 3 - 9 8 6 1 ( 0 2 ) 0 0 0 3 3 - 4
N. Swamy et al. / Archives of Biochemistry and Biophysics 402 (2002) 14–23
15
Fig. 1. Amino acid sequence of human DBP (Gc2 isomorph). Structural domains are identified according to albumin format. In the case of Gc1 isomorph, the amino acid sequence is the same except residues 152 is Glu, 311 is Arg, 416 is Glu, and 420 is Thr (marked by star symbols).
vitamin D3 (25-OH-D3 ); the latter in turn is transported (by DBP) to kidney to form 1a,25-dihydroxyvitamin D3 ð1; 25ðOHÞ2 D3 Þ, the active form of vitamin D hormone. Finally, 1,25ðOHÞ2 D3 is delivered to the target tissues by DBP. Thus biological functions of 1,25ðOHÞ2 D3 , viz. calcium and phosphorus homeostasis, immune modulation, regulation of growth, and maturity of normal and malignant cells, are intimately dependent on DBP. In addition to vitamin D sterol-binding, DBP plays an integral role in a circulating actin-scavenger system in plasma, in which plasma gelsolin severs the filaments of actin (F-actin) and DBP scavenges the monomeric actin (G-actin) with high efficiency [10–12]. It was also observed that this sequestration process is further aided by the rapid clearance of DBP–actin complex [13]. Thus, DBP prevents actin polymers from congesting arteries during cell injury and lysis. This function (of DBP) has profound implication in stroke, which is emphasized by the presence of DBP–actin complex in the serum of humans and animals sustaining injuries/inflammation,
e.g., trophoblastic emboli, severe active hepatitis, acute lung injury, etc. [14]. DBP is genetically related to serum albumin, alphafeto protein, and afamin [15–17]. It has a triple-domain modular structure, which is typical of the other members of this family [18]. Another characteristic (of the members of this protein family) is the presence of a large number of Cys residues, mostly in the oxidized state. In the case of DBP, twenty-eight (28) Cys residues (all in disulfide form) maintain its structural integrity (Fig. 1). During the past several years our laboratory and that of others have extensively studied the vitamin D sterolbinding properties of DBP by affinity/photoaffinity labeling with a goal to structurally define the vitamin D sterol-binding pocket of DBP [19–29]. These studies have demonstrated that vitamin D sterol- and actinbinding (by DBP) are restricted to the N-terminal (Domain I) and C-terminal (Domain II) areas of this protein, respectively. The purpose of the present investigation was to substantiate the results of the above-mentioned studies (to
16
N. Swamy et al. / Archives of Biochemistry and Biophysics 402 (2002) 14–23
structurally define vitamin D sterol-binding) with point mutation, truncation, and binding analysis, as well as to develop a DBP–actin crystal to study the relationship between vitamin D sterol- and actin-binding by DBP.
large number of colonies were found. The plasmid DNA was prepared from a few representative colonies and sequenced using automated DNA sequencer. Expression and purification of mutant DBP were carried out in a similar manner as that of the wild type, according to our published procedure [31] (vide infra).
Materials and methods Rabbit muscle acetone powder and 25-hydroxyvitamin D3 (25-OH-D3 ) were kind gifts from Dr. Andrew Szent-Gyorgyi, Department of Biological Sciences, Brandeis University, Waltham, MA, and Dr. Richard Gray, Amoco Research, Naperville, IL, respectively. 25-Hydroxy[26(27)-3 H]vitamin D3 (3 H-25-OH-D3 , sp. activity 20 Ci/mmol) was purchased from Amersham, Springfield, IL. Human plasma was obtained from American Red Cross, Dedham, MA. 25-OH-D3 –Sepharose affinity column (25-OH-D3 , covalently attached to Sepharose matrix) was prepared by a synthetic method developed in our laboratory [30]. DNA restriction and modification enzymes were purchased from New England Biolabs, Beverley, MA. Polymerase chain reactions were performed using the PCR kit purchased from Gibco BRL (Gaithersburg, MD) according to the manufacturer’s specifications. Oligonucleotides were synthesized on an Applied Biosystems DNA synthesizer, Foster City, CA (model ABI 394). Plasmid DNA sequencing was performed using Applied Biosystems ABI dye-terminator sequencing kit designed for use on Applied Biosystems automated sequencer (model ABI 373A). The amino acid sequence of reDBP was confirmed by N-terminal sequencing using an Applied Biosystems Peptide Sequencer. Peptide- and DNA-sequencing and DNA synthesis were carried out at the core peptide/DNA sequencing facility, Center for Advanced Biomedical Research, Boston University School of Medicine. Full-length recombinant hDBP (Gc2 isomorph) was expressed in Escherichia coli as a fusion partner of glutathione S-transferase (GST) according to our previously published procedure [31]. Site-directed mutagenesis of reDBP (Trp145–Ser145) Trp145 was changed to Ser by using a PCR-site directed mutagenesis. The TGG codon for Trp in the native cDNA was changed to TCC which encodes for Ser. The oligonucleotides that were used in the site-directed mutagenesis reaction were 50 AGUGGAAUAU UCGGAAUAAAUUGAUAAGCATATCC-30 and 50 AAUCCAUUUAUGTCCGA-AUAUUCCACUATT TACGGA-30 . The full-length cDNA in pEMBL 18 vector was used in the reaction. The PCR amplification, demethylation and uracil DNA-deglycosylation and cloning were performed as per manufacturer’s specifications. After transformation into E. coli host TB1, a
Expression of Domain I (1–203), Domain I–II (1–330) and Domain III (375–427) of hDBP Different structural domains of hDBP were expressed in E. coli similar to the full-length protein as GST fusion partners (as described earlier). DBP-Domain I was PCR amplified using 50 oligonucleotide with BamHI site: 50 TGC GGA TCC AGA GGC CGG GAT TAT GAA AAG-30 and 30 -oligonucleotide with XhoI site: 50 -TCG TTC TCG AGT TAC TAC ACT CTA TTT GAC AG30 with stop codons, and DBP-Domain III was PCR amplified with 50 -oligonucleotide with BamHI site: 50 CCC GGA TCC CTA CTA AAG AAG GAA CTA-30 and 30 -oligonucleotide with EcoRI site: 50 -GTT GAA TTC TCA GGA CTA CAG GAT ATT CTT-30 with stop codons. BamHI and XhoI sites were incorporated in the oligonucleotides (for 50 and 30 end, respectively) to facilitate directional cloning. The PCR products were digested with BamHI and XhoI restriction enzymes and cloned into pGEX-4T-2 expression vector. The constructs were designated as pGEX-DBP-Domain I and pGEX-DBP-Domain III. We used a different strategy to express Domain I + II. The cDNA encoding DBP (Gc2) contains an Acc I restriction site at position 330 located towards the end of Domain II. We digested pGEX-4T-2-DBP construct with AccI and XhoI which released all of the Domain III from the parent construct. Ligation after Klenow fill-in reaction produced a construct, which consisted of Domain I and most of Domain II (engineering of the restriction site resulted in incorporation of nine extra amino acids in the C-terminus). Different parts of DBP were expressed as GST-fusion proteins and purified in a procedure similar to that of full-length DBP. Briefly, TB1 cells carrying pGEX-DBP were grown to stationary phase at 37 °C in Luria broth (LB) containing 100 lg/ml ampicillin. The cultures were diluted 100 times with fresh LB-Amp and continued to grow until absorbance at 600 nm ðA600 Þ reached 0.4–0.6. The expression of fusion protein was induced by the addition of 0.5 mM IPTG for 12–14 h at 25 °C. Purification of GST–DBP fusion protein and recombinant DBP (reDBP) Purification of the GST–DBP fusion protein was performed according to the manufacturer’s specifications, with required modifications. The cells were lysed in TBS (50 mM Tris–HCl buffer, pH 8.3, 150 mM
N. Swamy et al. / Archives of Biochemistry and Biophysics 402 (2002) 14–23
NaCl), containing 1 mM phenylmethylsulfonyl fluoride (PMSF), 0.2 mg/ml lysozyme, and 0.1% Triton X-100, and disrupted by sonication. This was followed by removal of insoluble material by centrifugation at 12,000g for 30 min at 4 °C. The clear cell lysate, containing the fusion protein, was passed through glutathione–Agarose affinity column (Sigma Chemical, Milwaukee, WI), preequilibrated with TBS. The column was washed extensively with TBS to remove all unbound proteins, and columnbound GST–DBP was eluted with 20 mM oxidized glutathione in 100 mM Tris buffer, pH 8.3. The fusion protein was cleaved with thrombin (according to manufacturer’s specifications) and dialyzed against TBS to remove oxidized glutathione. Thrombin was removed by benzamidine–Sepharose chromatography (according to manufacturer’s specifications, Sigma). GST was separated from recombinant DBP (reDBP) using glutathione–Sepharose-4B affinity column. Sequence fidelities of the constructs (pET-DBP and pGEX-DBP) and reDBP were confirmed by DNA sequencing, and N-terminal peptide sequencing. Domain I (1–203), Domain I + II (1–330), and Domain III (375–427) of DBP had apparent molecular weights of 23, 38, and 10 kDa (determined by SDS– PAGE with molecular weight standards). Specific 3 H -25-hydroxyvitamin D3 -binding analysis of wild-type and mutant (Trp145–Ser145) DBP 3 H-25-OH-D3 -binding assays of wild and mutant varieties of reDBP were carried out by incubating protein samples (1.0 lg each, 19.6 pmol) in TEST buffer (50 mM Tris–HCl, 150 mM NaCl, 1.5 mM EDTA, 0.1% Triton X-100, pH 8.3) with 3 H-25-OH-D3 (3000 cpm, 0.15 pmol, dissolved in 10 ll EtOH) at 4 °C for 18 h followed by treatment with dextran-coated charcoal (at 4 °C), centrifugation, and radioactive counting. In each case, non-specific binding was determined by incubating the samples with 3 H-25-OH-D3 (3000 cpm, 0.15 pmol) and 25-OH-D3 (0:3 lg, 0.75 nmol), and specific binding was determined by subtracting non-specific binding from total binding. Each assay was carried out in triplicate.
Specific 3 H -25-hydroxyvitamin D3 -binding analysis of full-length reDBP and its structural domains (Domain I, Domain I + II, and Domain III) Samples of full-length reDBP and various domains (1:0 lg each, 19.6, 43.4, 26.3, and 100 pmol of native DBP, Domain I, Domain I + II, and Domain III, respectively) in Tris–HCl buffer, pH 8.3, were incubated with 3 H-25-OH-D3 (3000 cpm, 0.15 pmol, dissolved in 10 ll EtOH) at 4 °C for 18 h followed by treatment with dextran-coated charcoal (at 4 °C), centrifugation, and radioactive counting. In each case, non-specific binding was determined by incubating the samples with
17
3
H-25-OH-D3 (3000 cpm, 0.15 pmol) and 25-OHD3 ð0:3 lg, 0.75 nmol), and specific binding was determined by subtracting non-specific binding from total binding. Each assay was carried out in triplicate. Competitive binding assays of native DBP, full-length reDBP, and its structural domains (Domain I and Domain I + II) Competitive binding assays of native DBP, full-length reDBP, and its structural domains (Domain I and Domain I + II) with 3 H-25-OH-D3 were carried out according to published procedure [22]. Samples of native DBP, reDBP, and various domains (1:0 lg each) in Tris–HCl buffer, pH 8.3, were incubated with 3 H-25OH-D3 (2000 cpm, 0.1 pmol) and various doses of 25OH-D3 (0.92–62 pmol) for 20 h at 4 °C followed by incubation with dextran-coated charcoal (4 °C) for 15 min, centrifugation, and radioactive counting of the supernatants. Each assay was carried out in triplicate. Preparation of DBP–actin complex A crude preparation of actin was extracted by stirring 250 mg of rabbit muscle acetone powder in 10 ml of 10 mM Tris–HCl, pH 7.4 for 20 min on ice bath. The mixture was filtered through a syringe fitted with glass wool to clear the liquid from debris of muscle acetone powder. The clear liquid was added to 10 ml pooled human plasma maintained at 4 °C. The plasma–actin mixture was quickly adjusted to a final concentration of 10 mM Tris–HCl, pH 7.4, 0.1 mM EDTA, 0.5 mM ATP, and 0.2 mM CaCl2 using stock solutions, and the final solution was incubated at 4 °C for 60 min to allow the binding of actin to DBP in the plasma. Purification of DBP–actin complex was carried out by a modification of our procedure for the affinity purification of DBP from plasma [30]. Briefly, the mixture was loaded onto a 25-OH-D3 –Sepharose affinity column at 4 °C. The column was extensively washed with 10 mM Tris–HCl, pH 7.4, 0.1 mM EDTA, 0.5 mM ATP, 0.2 mM CaCl2 , and 0.1% Triton X-100 to remove the unbound proteins, and the bound DBP–actin complex was eluted using the same buffer containing 25-OH-D3 . The 25-OH-D3 eluted fraction from the affinity matrix was loaded onto a Bio Gel HTP column (BioRad, Richmond, CA) (2.0 ml), equilibrated with column buffer (10 mM potassium phosphate buffer, pH 7.0). First the column was washed with a large excess of column buffer (50 ml) to remove contaminants, and then the bound protein was eluted with 75 mM potassium phosphate buffer, pH 7.0. The protein was concentrated using Millipore spun protein concentrator (Millipore, Bedford, MA). The purity of the DBP–actin complex preparation was assessed by 10% SDS–PAGE alongside standard molecular weight marker proteins, purified DBP, and actin.
18
N. Swamy et al. / Archives of Biochemistry and Biophysics 402 (2002) 14–23
Competitive 3 H -25-OH-D3 -binding assays of DBP and DBP–actin complex Binding of 25-OH-D3 by DBP–actin complex was carried out according to the procedure described above. Briefly, samples of native DBP (affinity purified from human plasma) or DBP–actin complex (4.0 pmol each) in 0.5 ml 50 mM Tris–HCl, 150 mM NaCl, 1.5 mM EDTA, and 0.1% Triton X-100, pH 8.3, were incubated with 3 H-25-OH-D3 (0.139 pmol, in 10 ll EtOH) for 16 h at 4 °C in the absence or presence of increasing amounts of 25-OH-D3 (0.97–250 pmol). Following the incubation, the solutions were treated with ice-cold dextrancoated charcoal for 15 min at 4 °C and centrifuged at 5000g at 4 °C. The clear supernatants were mixed with scintillation cocktail and counted for radioactivity. Assays for each sample were carried out in triplicate. Stability of DBP–actin complex towards urea and trypsin The stability of the DBP–actin complex was investigated in the presence of different concentrations of urea. Thirty micrograms of DBP–actin complex was incubated with 0, 3.2, 4, and 5.3 M urea in Tris–HCl buffer, pH 7.4, for 30 min at 25 °C, followed by native-PAGE analysis. Similarly, stabilization of DBP–actin complex towards trypsin was determined by incubating samples of native DBP, actin, and DBP–actin complex in 50 mM Tris–HCl buffer, pH 7.4, with trypsin (20:1 protein:trypsin) for 14 h at 25 °C followed by analysis by 12% SDS–PAGE. Crystallization and X-ray diffraction analysis of DBP– actin complex For crystallization DBP–actin complex was prepared by mixing F-actin from rabbit muscle acetone powder with human DBP (Sigma) in a Tris buffer containing 5 mM ATP. DBP–actin complex was screened for suitable conditions for crystallization using hanging-drop method with commercially available screening kits (Hampton Research, CA). Based on these screens, platelike crystals were found to grow at 17 °C within 24 h in 0.1 M calcium acetate, 0.05 M cacodylate, pH 6.5, 9% PEG 8000, 10% glycerol, and 12 mg/ml DBP–actin. These crystals grew to dimensions of 0:4 0:4 0:05 mm over a period of several days. The crystals were flash frozen in a nitrogen stream at 100 K and X-ray diffraction data were collected at the Brookhaven National Synchrotron Light Source, beamline X8C.
Results and discussion Intense research has been carried out during the past several years to establish structure-functional basis of certain functions of DBP. For example, our group and
that of the late Dr. Haddad employed affinity/photoaffinity labeling technique to identify Domain I (of DBP) as the specific area for vitamin D sterol-binding [21,26,28]. Domain I contains Trp145, the only Trp residue in human DBP. We reported earlier that oxidation of this residue with N-bromosuccinimide led to a complete loss of 25-OH-D3 -binding, supporting the important role of Trp145 and Domain I towards vitamin D sterol-binding [32]. In the present study we substantiated the above chemical mutation data with site-specific mutation of Trp145 to Ser, and determined 25-OH-D3 -binding property of this mutant by ligand-binding assays. Traditionally vitamin D sterol (25-OH-D3 and other metabolites and synthetic analogs)-binding property of DBP is ascertained by competitive radioligand-binding assay of this protein. Previous studies have demonstrated that the dissociation constant between DBP and 25-OH-D3 (and several of its metabolites and synthetic analogs) is in the nanomolar range, suggesting a very slow dissociation of the holo-protein [33–35]. Therefore, in a competitive radioligand binding assay samples of DBP are incubated with a fixed quantity of 3 H-25-OHD3 , which is competed out from the binding pocket by various concentrations of vitamin D sterols or their synthetic analogs [20–27,33,34]. Specific activity of 3 H25-OH-D3 , used in these assays, is usually very high (Ci/ mmol range) to reduce non-specific binding and increase the sensitivity of competition. Furthermore, in a companion assay a large excess of 25-OH-D3 is added to compete out majority of 3 H-25-OH-D3 from the ligandbinding pocket. Radioactivity, associated with the latter DBP samples (non-specific binding), is subtracted from earlier mentioned samples to obtain ‘specific binding’. A similar procedure is also used in assaying 25-OH-D3 in clinical samples [33,34]. We have recently shown that 25-OH-D3 -binding characteristic of recombinant human DBP (Gc2 isomorph) is very similar to that of native DBP [31]. Therefore, we surmised that competitive binding assays could be used to determine 25-OH-D3 -binding property of the recombinant Trp145-Ser mutant. We carried out such an assay of native DBP and Trp145-Ser mutant to observe that the mutant had a low 25-OH-D3 -binding capability (results not shown). Therefore, we performed a 3 H-25-OH-D3 -binding-analysis, in which specific 3 H25-OH-D3 -binding of native DBP was compared with that of the mutant. As shown in Fig. 2, the mutant variety retained only approximately 20% of the specific 3 H-25-OH-D3 -binding of the wild-type. This result conclusively established the essential role of Trp145 towards vitamin sterol-binding, and further implicated Domain I as the area of DBP that is responsible for such activity. However, this information could not be extrapolated to ascertain whether other domains of DBP are also responsible for the binding of vitamin D ligands.
N. Swamy et al. / Archives of Biochemistry and Biophysics 402 (2002) 14–23
Fig. 2. Specific 3 H-25-OH-D3 -binding assays of reDBP and Trp145Ser mutant. Samples of wild-type and mutant DBP (1:0 lg each) were incubated with 3 H-25-OH-D3 (3000 cpm, 0.15 pmol, dissolved in 10 ll of EtOH) at 4 °C followed by removal of unbound 3 H-25-OH-D3 by charcoal treatment, centrifugation, and counting of radioactivity in the supernatants (total binding). Non-specific binding was determined by carrying out the same assays in the presence of an excess of 25-OH-D3 (0:3 lg, 0.75 nmol). Specific binding was calculated by subtracting non-specific binding from total binding in each case. Each assay was run in triplicate.
Fig. 3. SDS–PAGE analysis of various structural domains of DBP to show their homogeneity. Domain I: 1–203, Domain I + II: 1–330, Domain III: 375–427. Positions of standard molecular weight markers (MW) are shown on the right of each panel.
We expressed different structural domains of DBP in bacteria, and produced largely homogeneous samples of Domain I (1–203), Domain I + II (1–330), and Domain
19
III (375–427) as evidenced by SDS–PAGE-analysis (Fig. 3). As with the Trp145-Ser mutant, first we performed a specific 3 H-25-OH-D3 -binding assay of the domains with a single large dose of 25-OH-D3 to determine their capability to bind 3 H-25-OH-D3 in a specific manner, compared to full-length reDBP. These assays showed that Domain I (1–203), Domain I + II (1–330), and Domain III (375–427) retained approximately 88, 76.9, and 5.8%, respectively, of the specific 3 H-25-OH-D3 -binding by the full-length reDBP (100%) (results not shown). Therefore we performed competitive binding assays of native DBP, full-length reDBP, Domain I (1–203), and Domain I + II (1–330). As shown in Fig. 4, concentrations of 25-OH-D3 , displacing half-maximally bound 3 H-25-OH-D3 were 17.5, 11.4, 19, and 12.1 nM for native DBP, full-length reDBP, Domain I (1–203), and Domain I + II (1–330) respectively. Therefore, results of two assays (specific binding and competitive binding) in combination showed that Domain II, in addition to Domain I, might participate in vitamin D sterol-binding, while Domain III is largely not required for such binding activity. However, these results did not provide any conclusive proof whether Domain II is directly involved in 25-OHD3 -binding. It is probable that presence of Domain II stabilized Domain I by ‘correct’ conformational changes to promote 25-OH-D3 -binding. Other indirect effects also could not be ruled out. In our recent studies with affinity analogs of 25-OHD3 and 1,25ðOHÞ2 D3 containing the affinity probe at different parts of the parent vitamin D structure, we demonstrated that the vitamin D sterol-binding pocket of DBP is highly constrained, and does not allow any ‘tumbling’ of the ligand inside the binding pocket [26,27]. Results of the present studies further suggested that gross structural integrity of the protein is also required for specific ligand-binding. It is noteworthy that similar steric restrictions of the ligand-binding pocket are also observed in 1; 25ðOHÞ2 D3 -binding pocket of the vitamin D receptor (VDR), the nuclear receptor for 1; 25ðOHÞ2 D3 [36–38]. However, DBP and VDR are entirely different classes of proteins, and our findings suggested that these two proteins might have evolved from a primordial gene and separated during evolution. In 1992 Haddad et al. employed actin-affinity chromatography to study the binding of the proteolytic fragments of DBP and demonstrated that a C-terminal fragment of DBP (350–403), which spans most of Domain III, is capable of binding actin independently [28]. However, these studies could not determine any relationship between vitamin D and its metabolites and actin in terms of their binding by DBP. In an earlier study we coupled the 3-hydroxyl group of 25-OH-D3 to Sepharose through a 12 atom tether, and affinity-purified fully functional DBP from human serum [30]. We argued that if vitamin D sterol- and
20
N. Swamy et al. / Archives of Biochemistry and Biophysics 402 (2002) 14–23
Fig. 4. Competitive 3 H-25-OH-D3 -binding assays of native DBP, full-length reDBP, and its structural domains (Domain I and Domain I + II). Briefly, samples of native DBP, reDBP, and various domains (1.0 lg each) in Tris–HCl buffer, pH 8.3, were incubated with 3 H-25-OH-D3 (0.1 pmol) and various doses of 25-OH-D3 (0.92-62 pmol) for 20 h at 4 °C followed by incubation with dextran-coated charcoal (4 °C) for 15 min, centrifugation, and radioactive counting of the supernatants. In each case non-specific binding was determined by carrying out the same assay in the presence of an excess of 25-OH-D3 . Specific binding was calculated by subtracting non-specific binding from total binding. For each protein sample, percent of specific binding ðB=B0 100Þ is plotted against concentration of the competitor (25-OH-D3 ). Concentrations of 25-OH-D3 required to displace 50% of 3 H-25-OH-D3 bound to the protein samples are denoted on the Y-axis by an arrow.
Fig. 5. Cartoon depicting the formation of the complex between DBP in human plasma and actin in rabbit muscle, and binding of the complex to 25OH-D3 that is chemically attached to a solid matrix. The bound-protein could be removed from the affinity matrix by the natural ligand (25-OH-D3 ), and the purified DBP–actin complex could be obtained by removing bound-25-OH-D3 by hydroxylapatite chromatography. Please note that change in conformation of DBP upon binding to 25-OH-D3 (chemically bound to matrix and free) is represented by slight change of its shape and graphic pattern.
actin-binding by DBP are largely independent processes, we might be able to employ this matrix to affinity purify a DBP–actin complex from human plasma, as depicted in Fig. 5.
Preparation of DBP–actin complex is described in the Materials and methods section. The denaturing SDS–PAGE analysis of the protein fraction, eluted with 25-OH-D3 from the affinity matrix, showed that the
N. Swamy et al. / Archives of Biochemistry and Biophysics 402 (2002) 14–23
21
Fig. 6. SDS–PAGE and native-PAGE analyses of DBP, actin, and DBP–actin complex. Positions of the standard molecular weight markers are denoted on the right of the left panel.
protein fraction was homogeneous, and contained only DBP and actin (Fig. 6, left panel). On the other hand, native PAGE analysis revealed a single band moving slower than DBP or actin (Fig. 6, right panel), proving the compositional homogeneity of the DBP–actin complex. In essence, we generated DBP–actin complex in plasma, and purified the complex by using 25-OHD3 –Sepharose affinity chromatography. These results suggested that the binding of vitamin D sterols (by DBP) might not influence the binding of actin in a significant manner. The above observation was further supported by the 3 H-25-OH-D3 -binding assays of DBP and DBP–actin
complex. As shown in Fig. 7, incubation of DBP or DBP–actin complex with 3 H-25-OH-D3 (0.139 pmol) in the presence of increasing amounts of 25-OH-D3 (0.97– 250 pmol) resulted in a dose-dependent displacement of 3 H-25-OH-D3 from the 25-OH-D3 -binding pocket of DBP in both cases. Additionally, the competitive binding curves were almost identical indicating that 25-OHD3 -binding characteristics of DBP–actin complex are very similar to those of DBP alone. McLeod et al. [39] demonstrated that hDBP, bound to 25-OH-D3 and 1,25ðOHÞ2 D3 , bound actin in a manner that was indistinguishable from apo-hDBP. Therefore, this information, in combination with the
Fig. 7. Competitive 3 H-25-OH-D3 -binding assays of native DBP and DBP–actin complex. Briefly, samples (0.2 lg, 4.0 pmol each) of native DBP (affinity-purified from human plasma) or DBP–actin complex (from human plasma and rabbit muscle) in 50 mM Tris–HCl, 150 mM NaCl, 1.5 mM of EDTA, and 0.1% Triton X-100, pH 8.3 were incubated with 3 H-25-OH-D3 (0.139 pmol) for 16 h at 4 °C in the absence or presence of various amounts of 25-OH-D3 (0.97–250 pmol). Following the incubation, the solutions were treated with ice-cold dextran-coated charcoal for 15 min at 4 °C and centrifuged at 5000g at 4 °C. The supernatants were mixed with scintillation cocktail and counted for radioactivity.
22
N. Swamy et al. / Archives of Biochemistry and Biophysics 402 (2002) 14–23
results described in this communication strongly suggested that vitamin D sterol- and actin-binding by DBP might be independent processes to a large extent. However, it should be emphasized that these results do not conclusively prove total independence of these binding processes. We argued that proof of our above-mentioned structure-functional analysis could be best achieved by determining the three-dimensional structure of DBP– actin by X-ray crystallography, and studying its prop erty in the presence of 25-OH-D3 . Recently a 2.3 A crystal structure of DBP–25-OH-D3 complex was reported, which revealed that the vitamin D-binding site is in the N-terminal part of the protein [40–42]. This communication also reports significant difference in the overall structure, as well as local folding pattern between the structures of DBP and serum albumin, a homologous protein, limiting the scope of structure determination of DBP–actin complex by molecular replacement technique based on the crystal structure of G-actin [43,44]. Prior to attempting to crystallize DBP–actin complex it was important to determine its stability under various conditions. Susceptibility of a protein in the apo- and holo-forms towards trypsin treatment is often used to determine its conformational changes upon ligandbinding. As shown in Fig. 8, lane 1, trypsinization of DBP primarily resulted in a prominent 25 kDa band and several low molecular weight peptides moving with the dye front. The amino acid sequence analysis of this 25 kDa trypsin-resistant band revealed that it belonged to
Fig. 8. Tryptic digestion of DBP, actin, and DBP–actin. Samples of native DBP, actin, and DBP–actin complex (30 lg each) in 50 mM Tris–HCl buffer, pH 7.4 were incubated with trypsin (20:1 protein:trypsin) for 14 h at 25 °C followed by analysis with 12% SDS– PAGE. Lane 1: DBP + trypsin; Lane 2: DBP–actin complex + trypsin; Lane 3: actin + trypsin; Lane 4: DBP–actin complex; MW: standard molecular weight markers (positions are shown on the right).
Domain I with five (5) amino acids truncated in the Nterminus. This indicated that while Domains II and III are trypsin susceptible, Domain I is largely resistant to trypsin cleavage. In contrast, trypsinization of actin proteolyzed the protein completely resulting in several low molecular weight peptides (Fig. 8, lane 3). On the other hand, trypsin treatment of DBP–actin complex showed two prominent bands of 52 and 34 kDa and other lower molecular weight peptides (Fig. 8, lane 2). Amino acid sequencing of the 52 and 34 kDa bands revealed that the 52 kDa band was DBP (5 amino acids shorter in Nterminus), and 34 kDa originated from actin (starting with amino acid position 69). These results indicated that DBP, particularly its Domains II and III were completely protected from trypsin-cleavage upon actinbinding, suggesting that Domain II, in addition to Domain III, might be involved in actin-binding. Similarly, actin, except its N-terminus, was also protected from tryptic digestion when complexed with DBP. Apparently, the interaction between DBP and actin stabilized both proteins and masked their trypsin sensitive sites. We also observed that DBP–actin complex was stable for at least 60 days at 4 °C. Furthermore, the complex was stable in the presence of up to 4.0 M urea (there was only partial dissociation in 5.3 M urea) at 25 °C as determined by native PAGE (results not shown). Observed stability of DBP–actin complex provided the basis for its crystallization studies, which were carried out according to the method described in the Matrials and methods section. Diffraction data collected from crystals were processed with DENZO/SCALEPACK [45] and showed the space group to be primitive , orthorhombic with unit cell dimensions a ¼ 80:2 A , and c ¼ 159:6 A . A full data set was colb ¼ 87:3 A . Systematic absences show the space lected to 2.5 A group to be P21 21 21 and Vm calculations indicated a single molecule of the complex in the asymmetric unit (Vm ¼ 2:9). The availability of crystals of the DBP–actin complex set the stage for studying its interaction with 25-OH-D3 . But, we have so far been unable to grow crystals of actin– DBP in the presence of 25-OH-D3 . Soaking of preexisting crystals of actin–DBP in crystallization buffer containing various concentrations of 25-OH-D3 resulted in cracking of the crystal. We have demonstrated that DBP–actin bound 25-OH-D3 similar to apo-DBP (Fig. 7). Conversely, McLeod et al. showed that DBP–25-OHD3 holo-protein bound DBP in a manner indistinguishable from apo-DBP [39]. Therefore, crystal cracking might reflect a ligand-induced conformational change in the complex, disrupting crystal contacts. However, a conclusive proof of this phenomenon would require knowledge of the three-dimensional structure of the DBP–actin complex. Currently structure determination is being pursued in our laboratory by a combination of
N. Swamy et al. / Archives of Biochemistry and Biophysics 402 (2002) 14–23
molecular replacement methods and heavy metal derivatization. Molecular replacement has provided a solution for the position of the actin in the complex, but the phases are inadequate for tracing the chain of the DBP. Molecular replacement using models based on the known structure of albumin, a DBP homolog, has not provided a solution. We are therefore working to obtain additional phase information from heavy metal derivatives using either isomorphous replacement or multiwavelength anomalous diffraction methods. In conclusion, we have probed the vitamin D steroland actin-binding properties of DBP with various techniques to establish structural basis for the multipleligand-binding properties of DBP. We have shown for the first time that structural domains I and II of DBP are required for vitamin D sterol-binding. We have also shown that in solution vitamin D sterol- and actinbinding are largely independent processes, although the binding of 25-OH-D3 likely results in conformational changes in the complex.
Acknowledgments The work was supported by a grant from the National Institute of Digestive, Diabetes and Kidney Diseases (#DK 44337 to R.R.). Assistance from Dr. Jan Krzyszt of Blusztajn in the preparation of the manuscript is gratefully acknowledged. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13]
N.E. Cooke, J.G. Haddad, Endocr. Rev. 4 (1996) 125–128. R. Ray, Proc. Soc. Exp. Biol. Med. 212 (1996) 305–312. J.G. Haddad, J. Steroid Biochem. Mol. Biol. 53 (1995) 579–582. N. Yamamoto, S. Homma, Proc. Natl. Acad. Sci. 88 (1991) 8539– 8543. N. Yamamoto, S. Homma, I. Millman, J. Immunol. 147 (1991) 273–280. G.B. Schneider, K.A. Benis, N.W. Flay, R.A. Ireland, S.N. Popoff, Bone 16 (1995) 657–662. N. Swamy, S. Ghosh, G.B. Schneider, R. Ray, J. Cell Biochem. 81 (2001) 535–546. N. Yamammoto, V.R. Naraparaju, Cancer Res. 57 (1997) 2187– 2192. Y. Koga, V.R. Naraparaju, N. Yamamoto, Proc. Soc. Exp. Biol. Med. 220 (1999) 20–26. H. Van Baelen, R. Bouillon, P. DeMoor, J. Biol. Chem. 255 (1980) 2270–2272. J.G. Haddad, Arch. Biochem. Biophys. 213 (1982) 538–544. P.J. Goldschmidt-Clermont, M.H. Williams, R.M. Galbraith, Biochem. Biophys. Res. Commun. 146 (1987) 611–617. S.E. Lind, D.B. Smith, P.A. Janmey, T.P. Stossel, J. Clin. Invest. 78 (1986) 736–742.
23
[14] W. Lee, R.M. Galbraith, N. Engl. J. Med. 326 (1992) 1335– 1341. [15] F. Yang, J.I. Brune, S.L. Naylor, R.L. Cupples, K.H. Naberhaus, B.H. Bowman, Proc. Natl. Acad. Sci. USA 82 (1985) 7994–7998. [16] N.E. Cooke, E.V. David, J. Clin. Invest. 76 (1985) 2420–2424. [17] H.S. Lichenstein, D.E. Lyons, M.M. Wurfel, D.A. Johnson, M.D. McGinley, J.C. Leidli, D.B. Trollinger, J.P. Mayer, S.D. Wright, M.M.. Zukowski, J. Biol. Chem. 269 (1994) 18149–18154. [18] F. Schoentgen, M. Metz-Boutigue, J. Jolles, J. Constans, P. Jolles, Biochem. Biophys. Acta 871 (1986) 189–198. [19] R. Ray, S.A. Holick, N. Hanafin, M.F. Holick, Biochemistry 25 (1986) 4729–4733. [20] R. Ray, R. Bouillon, H.G. Baelen, M.F. Holick, Biochemistry 36 (1991) 4809–4813. [21] R. Ray, R. Bouillon, H.G. Van Baelen, M.F. Holick, Biochemistry 30 (1991) 7638–7642. [22] N. Swamy, R. Ray, Arch. Biochem. Biophys. 319 (1995) 504– 507. [23] N. Swamy, R. Ray, Arch. Biochem. Biophys. 333 (1996) 139–144. [24] J.K. Addo, R. Ray, Steroids 63 (1998) 218–223. [25] J.K. Addo, N. Swamy, R. Ray, Steroids 64 (1999) 273–282. [26] N. Swamy, J.K. Addo, M.R. Uskokovic, R. Ray, Arch. Biochem. Biophys. 373 (2000) 471–478. [27] N. Swamy, A. Dutta, R. Ray, Biochemistry 36 (1997) 7432–7436. [28] J.G. Haddad, Y.Z. Hu, M.A. Kowalski, C. Laramore, K. Ray, P. Robzyk, N.E. Cooke, Biochemistry 31 (1992) 7174–7181. [29] R. Ray, in: M.F. Holick (Ed.), Vitamin D: Physiology, Molecular Biology and Clinical Applications, Humana Press, New Jersey, 1999, pp. 147–162. [30] N. Swamy, A. Roy, R. Chang, M. Brisson, R. Ray, Protein Expr. Purif. 6 (1995) 185–188. [31] N. Swamy, S. Ghosh, R. Ray, Protein Expr. Purif. 10 (1997) 115– 122. [32] N. Swamy, M. Brisson, R. Ray, J. Biol Chem. 270 (1995) 2636– 2639. [33] J.G. Hadadad, Methods Enzymol. 67 (1980) 449–459. [34] T.C. Chen, A.K. Turner, M.F. Holick, J. Nutr. Biochem. 1 (1990) 315–319. [35] R. Bouillon, W.H. Okamura, A.W. Norman, Endocr. Rev. 16 (1995) 200–257. [36] N. Rochel, J.M. Wurtz, A. Mitschler, B. Klaholz, D. Moras, Mol. Cell 5 (2000) 173–179. [37] N. Swamy, W. Xu, N. Paz, J-C. Hsieh, M.R. Haussler, G.J. Maalouf, S.C. Mohr, R. Ray, Biochemistry 39 (2000) 12162– 12171. [38] S.C. Mohr, N. Swamy, W. Xu, R. Ray, Steroids 66 (2001) 189– 201. [39] J.F. McLeod, M.A. Kowalski, J.G. Haddad, J. Biol. Chem. 264 (1989) 1260–1267. [40] C. Verboven, A. Rabijns, M. De Maeyer, H. Van Baelen, R. Bouillon, C. De Ranter, Nature Struct. Biol. 2 (2002) 131–136. [41] N.J. Vogelaar, U. Lindberg, C.E. Schutt, J. Mol. Biol. 220 (1991) 545–547. [42] C.C. Verboven, H.L. De Bondt, C. De Ranter, R. Bouillon, H. Van Baelen, J. Steroid Biochem. Mol. Biol. 54 (1995) 11–14. [43] W. Kabsch, H.G. Mannherz, D. Suck, E.F. Pai, K.C. Holmes, Nature 347 (1990) 37–44. [44] L.R. Otterbein, P. Graceffa, R. Dominguez, Science 293 (2001) 616–618. [45] W. Minor, D. Tomchick, Z. Otwinowski, Struct. Fold Des. 8 (2000) R105–110.
Biochemical and preliminary crystallographic characterization of the vitamin D sterol- and actin-binding by human vitamin D-binding protein Narasimha Swamy,a James F. Head,b Daniel Weitz,a and Rahul Raya,* a
Bioorganic Chemistry & Structural Biology, Section in Endocrinology, Diabetes and Metabolism, Department of Medicine, Boston University School of Medicine, Boston, MA 02118, USA b Department of Physiology & Biophysics, Boston University School of Medicine, Boston, MA 02118, USA Received 20 December 2001, and in revised form 19 February 2002
Abstract Vitamin D-binding protein (DBP), a multi-functional serum glycoprotein, has a triple-domain modular structure. Mutation of Trp145 (in Domain I) to Ser decreased 25-OH-D3 -binding by 80%. Furthermore, recombinant Domain I (1–203) and Domain I + II (1–330) showed specific and strong binding for 25-OH-D3 , but Domain III (375–427) did not, suggesting that only Domains I and II might be required for vitamin D sterol-binding. Past studies have suggested that Domain III is independently capable of binding G-actin. We exploited this apparently independent ligand-binding property of DBP to purify DBP–actin complex from human serum and rabbit muscle actin by 25-OH-D3 affinity chromatography. Competitive 3 H-25-OH-D3 binding curves for native DBP and DBP–actin complex were almost identical, further suggesting that vitamin D sterol- and actin-binding activities by DBP might be largely independent of each other. Trypsin treatment of DBP produced a prominent 25 kDa band (Domain I, minus 5 amino acids in N-terminus), while actin was completely fragmented by such treatment. In contrast, tryptic digestion of purified DBP–actin complex showed two prominent bands, 52 (DBP, minus 5 amino acids in the N-terminus) and 34 kDa (actin, starting with amino acid position 69) indicating that DBP, particularly its Domains II and III were protected from trypsin cleavage upon actin-binding. Similarly, actin, except its N-terminus, was also protected from tryptic digestion when complexed with DBP. These results provided the basis crystal, primitive orthorhombic with unit cell dimensions for our studies to crystallize DBP–actin complex, which produced a 2.5 A , b ¼ 87:3 A , and c ¼ 159:6 A , P21 21 21 space group, Vm ¼ 2:9. Soaking of crystals of actin–DBP in crystallization buffer a ¼ 80:2 A containing various concentrations of 25-OH-D3 resulted in cracking of the crystal, which was probably a reflection of a ligandinduced conformational change in the complex, disrupting crystal contacts. In conclusion, we have provided data to suggest that although binding of 25-OH-D3 to DBP might result in discrete conformational changes in the holo-protein to influence actin-binding, these binding processes are largely independent of each other in solution. Ó 2002 Elsevier Science (USA). All rights reserved. Keywords: Vitamin D-binding protein; G-actin; 25-hydroxyvitamin D3 ; 1,25-dihydroxyvitamin D3 ; DBP-mutant; Recombinant DBP-structural domains; DBP–actin crystal
Vitamin D-binding protein (DBP)1 is a polymorphic serum glycoprotein with multiple functions that include highly specific binding of vitamin D sterols, G-actin, fatty acids, and chemotactic agents [1–3]. Furthermore, *
Corresponding author. Fax: +617-638-8194. E-mail address: [email protected] (R. Ray). 1 Abbreviations used: hDBP, human vitamin D-binding protein; DBP-maf, DBP-macrophage-activating factor; 25-OH-D3 , 25-hydroxyvitamin D3 ; 1; 25ðOHÞ2 D3 , la,25-dihydroxyvitamin D3 ; reDBP, recombinant DBP.
DBP undergoes a sequential removal of sugar moieties (by glycosidases produced by inflammation-primed B- and T-lymphocytes) to generate a posttranslationally modified form of DBP (DBP-macrophage-activating factor, DBP-maf), which is a potent activator of macrophages and osteoclasts [4–7]. Furthermore, DBP-maf was found to increase the survival rate of mice bearing Ehrlich ascites tumor [8,9]. Vitamin D sterol transport by DBP has been studied extensively [1–3]. DBP binds vitamin D3 in serum and transports this seco-steroid to liver to form 25-hydroxy-
0003-9861/02/$ - see front matter Ó 2002 Elsevier Science (USA). All rights reserved. PII: S 0 0 0 3 - 9 8 6 1 ( 0 2 ) 0 0 0 3 3 - 4
N. Swamy et al. / Archives of Biochemistry and Biophysics 402 (2002) 14–23
15
Fig. 1. Amino acid sequence of human DBP (Gc2 isomorph). Structural domains are identified according to albumin format. In the case of Gc1 isomorph, the amino acid sequence is the same except residues 152 is Glu, 311 is Arg, 416 is Glu, and 420 is Thr (marked by star symbols).
vitamin D3 (25-OH-D3 ); the latter in turn is transported (by DBP) to kidney to form 1a,25-dihydroxyvitamin D3 ð1; 25ðOHÞ2 D3 Þ, the active form of vitamin D hormone. Finally, 1,25ðOHÞ2 D3 is delivered to the target tissues by DBP. Thus biological functions of 1,25ðOHÞ2 D3 , viz. calcium and phosphorus homeostasis, immune modulation, regulation of growth, and maturity of normal and malignant cells, are intimately dependent on DBP. In addition to vitamin D sterol-binding, DBP plays an integral role in a circulating actin-scavenger system in plasma, in which plasma gelsolin severs the filaments of actin (F-actin) and DBP scavenges the monomeric actin (G-actin) with high efficiency [10–12]. It was also observed that this sequestration process is further aided by the rapid clearance of DBP–actin complex [13]. Thus, DBP prevents actin polymers from congesting arteries during cell injury and lysis. This function (of DBP) has profound implication in stroke, which is emphasized by the presence of DBP–actin complex in the serum of humans and animals sustaining injuries/inflammation,
e.g., trophoblastic emboli, severe active hepatitis, acute lung injury, etc. [14]. DBP is genetically related to serum albumin, alphafeto protein, and afamin [15–17]. It has a triple-domain modular structure, which is typical of the other members of this family [18]. Another characteristic (of the members of this protein family) is the presence of a large number of Cys residues, mostly in the oxidized state. In the case of DBP, twenty-eight (28) Cys residues (all in disulfide form) maintain its structural integrity (Fig. 1). During the past several years our laboratory and that of others have extensively studied the vitamin D sterolbinding properties of DBP by affinity/photoaffinity labeling with a goal to structurally define the vitamin D sterol-binding pocket of DBP [19–29]. These studies have demonstrated that vitamin D sterol- and actinbinding (by DBP) are restricted to the N-terminal (Domain I) and C-terminal (Domain II) areas of this protein, respectively. The purpose of the present investigation was to substantiate the results of the above-mentioned studies (to
16
N. Swamy et al. / Archives of Biochemistry and Biophysics 402 (2002) 14–23
structurally define vitamin D sterol-binding) with point mutation, truncation, and binding analysis, as well as to develop a DBP–actin crystal to study the relationship between vitamin D sterol- and actin-binding by DBP.
large number of colonies were found. The plasmid DNA was prepared from a few representative colonies and sequenced using automated DNA sequencer. Expression and purification of mutant DBP were carried out in a similar manner as that of the wild type, according to our published procedure [31] (vide infra).
Materials and methods Rabbit muscle acetone powder and 25-hydroxyvitamin D3 (25-OH-D3 ) were kind gifts from Dr. Andrew Szent-Gyorgyi, Department of Biological Sciences, Brandeis University, Waltham, MA, and Dr. Richard Gray, Amoco Research, Naperville, IL, respectively. 25-Hydroxy[26(27)-3 H]vitamin D3 (3 H-25-OH-D3 , sp. activity 20 Ci/mmol) was purchased from Amersham, Springfield, IL. Human plasma was obtained from American Red Cross, Dedham, MA. 25-OH-D3 –Sepharose affinity column (25-OH-D3 , covalently attached to Sepharose matrix) was prepared by a synthetic method developed in our laboratory [30]. DNA restriction and modification enzymes were purchased from New England Biolabs, Beverley, MA. Polymerase chain reactions were performed using the PCR kit purchased from Gibco BRL (Gaithersburg, MD) according to the manufacturer’s specifications. Oligonucleotides were synthesized on an Applied Biosystems DNA synthesizer, Foster City, CA (model ABI 394). Plasmid DNA sequencing was performed using Applied Biosystems ABI dye-terminator sequencing kit designed for use on Applied Biosystems automated sequencer (model ABI 373A). The amino acid sequence of reDBP was confirmed by N-terminal sequencing using an Applied Biosystems Peptide Sequencer. Peptide- and DNA-sequencing and DNA synthesis were carried out at the core peptide/DNA sequencing facility, Center for Advanced Biomedical Research, Boston University School of Medicine. Full-length recombinant hDBP (Gc2 isomorph) was expressed in Escherichia coli as a fusion partner of glutathione S-transferase (GST) according to our previously published procedure [31]. Site-directed mutagenesis of reDBP (Trp145–Ser145) Trp145 was changed to Ser by using a PCR-site directed mutagenesis. The TGG codon for Trp in the native cDNA was changed to TCC which encodes for Ser. The oligonucleotides that were used in the site-directed mutagenesis reaction were 50 AGUGGAAUAU UCGGAAUAAAUUGAUAAGCATATCC-30 and 50 AAUCCAUUUAUGTCCGA-AUAUUCCACUATT TACGGA-30 . The full-length cDNA in pEMBL 18 vector was used in the reaction. The PCR amplification, demethylation and uracil DNA-deglycosylation and cloning were performed as per manufacturer’s specifications. After transformation into E. coli host TB1, a
Expression of Domain I (1–203), Domain I–II (1–330) and Domain III (375–427) of hDBP Different structural domains of hDBP were expressed in E. coli similar to the full-length protein as GST fusion partners (as described earlier). DBP-Domain I was PCR amplified using 50 oligonucleotide with BamHI site: 50 TGC GGA TCC AGA GGC CGG GAT TAT GAA AAG-30 and 30 -oligonucleotide with XhoI site: 50 -TCG TTC TCG AGT TAC TAC ACT CTA TTT GAC AG30 with stop codons, and DBP-Domain III was PCR amplified with 50 -oligonucleotide with BamHI site: 50 CCC GGA TCC CTA CTA AAG AAG GAA CTA-30 and 30 -oligonucleotide with EcoRI site: 50 -GTT GAA TTC TCA GGA CTA CAG GAT ATT CTT-30 with stop codons. BamHI and XhoI sites were incorporated in the oligonucleotides (for 50 and 30 end, respectively) to facilitate directional cloning. The PCR products were digested with BamHI and XhoI restriction enzymes and cloned into pGEX-4T-2 expression vector. The constructs were designated as pGEX-DBP-Domain I and pGEX-DBP-Domain III. We used a different strategy to express Domain I + II. The cDNA encoding DBP (Gc2) contains an Acc I restriction site at position 330 located towards the end of Domain II. We digested pGEX-4T-2-DBP construct with AccI and XhoI which released all of the Domain III from the parent construct. Ligation after Klenow fill-in reaction produced a construct, which consisted of Domain I and most of Domain II (engineering of the restriction site resulted in incorporation of nine extra amino acids in the C-terminus). Different parts of DBP were expressed as GST-fusion proteins and purified in a procedure similar to that of full-length DBP. Briefly, TB1 cells carrying pGEX-DBP were grown to stationary phase at 37 °C in Luria broth (LB) containing 100 lg/ml ampicillin. The cultures were diluted 100 times with fresh LB-Amp and continued to grow until absorbance at 600 nm ðA600 Þ reached 0.4–0.6. The expression of fusion protein was induced by the addition of 0.5 mM IPTG for 12–14 h at 25 °C. Purification of GST–DBP fusion protein and recombinant DBP (reDBP) Purification of the GST–DBP fusion protein was performed according to the manufacturer’s specifications, with required modifications. The cells were lysed in TBS (50 mM Tris–HCl buffer, pH 8.3, 150 mM
N. Swamy et al. / Archives of Biochemistry and Biophysics 402 (2002) 14–23
NaCl), containing 1 mM phenylmethylsulfonyl fluoride (PMSF), 0.2 mg/ml lysozyme, and 0.1% Triton X-100, and disrupted by sonication. This was followed by removal of insoluble material by centrifugation at 12,000g for 30 min at 4 °C. The clear cell lysate, containing the fusion protein, was passed through glutathione–Agarose affinity column (Sigma Chemical, Milwaukee, WI), preequilibrated with TBS. The column was washed extensively with TBS to remove all unbound proteins, and columnbound GST–DBP was eluted with 20 mM oxidized glutathione in 100 mM Tris buffer, pH 8.3. The fusion protein was cleaved with thrombin (according to manufacturer’s specifications) and dialyzed against TBS to remove oxidized glutathione. Thrombin was removed by benzamidine–Sepharose chromatography (according to manufacturer’s specifications, Sigma). GST was separated from recombinant DBP (reDBP) using glutathione–Sepharose-4B affinity column. Sequence fidelities of the constructs (pET-DBP and pGEX-DBP) and reDBP were confirmed by DNA sequencing, and N-terminal peptide sequencing. Domain I (1–203), Domain I + II (1–330), and Domain III (375–427) of DBP had apparent molecular weights of 23, 38, and 10 kDa (determined by SDS– PAGE with molecular weight standards). Specific 3 H -25-hydroxyvitamin D3 -binding analysis of wild-type and mutant (Trp145–Ser145) DBP 3 H-25-OH-D3 -binding assays of wild and mutant varieties of reDBP were carried out by incubating protein samples (1.0 lg each, 19.6 pmol) in TEST buffer (50 mM Tris–HCl, 150 mM NaCl, 1.5 mM EDTA, 0.1% Triton X-100, pH 8.3) with 3 H-25-OH-D3 (3000 cpm, 0.15 pmol, dissolved in 10 ll EtOH) at 4 °C for 18 h followed by treatment with dextran-coated charcoal (at 4 °C), centrifugation, and radioactive counting. In each case, non-specific binding was determined by incubating the samples with 3 H-25-OH-D3 (3000 cpm, 0.15 pmol) and 25-OH-D3 (0:3 lg, 0.75 nmol), and specific binding was determined by subtracting non-specific binding from total binding. Each assay was carried out in triplicate.
Specific 3 H -25-hydroxyvitamin D3 -binding analysis of full-length reDBP and its structural domains (Domain I, Domain I + II, and Domain III) Samples of full-length reDBP and various domains (1:0 lg each, 19.6, 43.4, 26.3, and 100 pmol of native DBP, Domain I, Domain I + II, and Domain III, respectively) in Tris–HCl buffer, pH 8.3, were incubated with 3 H-25-OH-D3 (3000 cpm, 0.15 pmol, dissolved in 10 ll EtOH) at 4 °C for 18 h followed by treatment with dextran-coated charcoal (at 4 °C), centrifugation, and radioactive counting. In each case, non-specific binding was determined by incubating the samples with
17
3
H-25-OH-D3 (3000 cpm, 0.15 pmol) and 25-OHD3 ð0:3 lg, 0.75 nmol), and specific binding was determined by subtracting non-specific binding from total binding. Each assay was carried out in triplicate. Competitive binding assays of native DBP, full-length reDBP, and its structural domains (Domain I and Domain I + II) Competitive binding assays of native DBP, full-length reDBP, and its structural domains (Domain I and Domain I + II) with 3 H-25-OH-D3 were carried out according to published procedure [22]. Samples of native DBP, reDBP, and various domains (1:0 lg each) in Tris–HCl buffer, pH 8.3, were incubated with 3 H-25OH-D3 (2000 cpm, 0.1 pmol) and various doses of 25OH-D3 (0.92–62 pmol) for 20 h at 4 °C followed by incubation with dextran-coated charcoal (4 °C) for 15 min, centrifugation, and radioactive counting of the supernatants. Each assay was carried out in triplicate. Preparation of DBP–actin complex A crude preparation of actin was extracted by stirring 250 mg of rabbit muscle acetone powder in 10 ml of 10 mM Tris–HCl, pH 7.4 for 20 min on ice bath. The mixture was filtered through a syringe fitted with glass wool to clear the liquid from debris of muscle acetone powder. The clear liquid was added to 10 ml pooled human plasma maintained at 4 °C. The plasma–actin mixture was quickly adjusted to a final concentration of 10 mM Tris–HCl, pH 7.4, 0.1 mM EDTA, 0.5 mM ATP, and 0.2 mM CaCl2 using stock solutions, and the final solution was incubated at 4 °C for 60 min to allow the binding of actin to DBP in the plasma. Purification of DBP–actin complex was carried out by a modification of our procedure for the affinity purification of DBP from plasma [30]. Briefly, the mixture was loaded onto a 25-OH-D3 –Sepharose affinity column at 4 °C. The column was extensively washed with 10 mM Tris–HCl, pH 7.4, 0.1 mM EDTA, 0.5 mM ATP, 0.2 mM CaCl2 , and 0.1% Triton X-100 to remove the unbound proteins, and the bound DBP–actin complex was eluted using the same buffer containing 25-OH-D3 . The 25-OH-D3 eluted fraction from the affinity matrix was loaded onto a Bio Gel HTP column (BioRad, Richmond, CA) (2.0 ml), equilibrated with column buffer (10 mM potassium phosphate buffer, pH 7.0). First the column was washed with a large excess of column buffer (50 ml) to remove contaminants, and then the bound protein was eluted with 75 mM potassium phosphate buffer, pH 7.0. The protein was concentrated using Millipore spun protein concentrator (Millipore, Bedford, MA). The purity of the DBP–actin complex preparation was assessed by 10% SDS–PAGE alongside standard molecular weight marker proteins, purified DBP, and actin.
18
N. Swamy et al. / Archives of Biochemistry and Biophysics 402 (2002) 14–23
Competitive 3 H -25-OH-D3 -binding assays of DBP and DBP–actin complex Binding of 25-OH-D3 by DBP–actin complex was carried out according to the procedure described above. Briefly, samples of native DBP (affinity purified from human plasma) or DBP–actin complex (4.0 pmol each) in 0.5 ml 50 mM Tris–HCl, 150 mM NaCl, 1.5 mM EDTA, and 0.1% Triton X-100, pH 8.3, were incubated with 3 H-25-OH-D3 (0.139 pmol, in 10 ll EtOH) for 16 h at 4 °C in the absence or presence of increasing amounts of 25-OH-D3 (0.97–250 pmol). Following the incubation, the solutions were treated with ice-cold dextrancoated charcoal for 15 min at 4 °C and centrifuged at 5000g at 4 °C. The clear supernatants were mixed with scintillation cocktail and counted for radioactivity. Assays for each sample were carried out in triplicate. Stability of DBP–actin complex towards urea and trypsin The stability of the DBP–actin complex was investigated in the presence of different concentrations of urea. Thirty micrograms of DBP–actin complex was incubated with 0, 3.2, 4, and 5.3 M urea in Tris–HCl buffer, pH 7.4, for 30 min at 25 °C, followed by native-PAGE analysis. Similarly, stabilization of DBP–actin complex towards trypsin was determined by incubating samples of native DBP, actin, and DBP–actin complex in 50 mM Tris–HCl buffer, pH 7.4, with trypsin (20:1 protein:trypsin) for 14 h at 25 °C followed by analysis by 12% SDS–PAGE. Crystallization and X-ray diffraction analysis of DBP– actin complex For crystallization DBP–actin complex was prepared by mixing F-actin from rabbit muscle acetone powder with human DBP (Sigma) in a Tris buffer containing 5 mM ATP. DBP–actin complex was screened for suitable conditions for crystallization using hanging-drop method with commercially available screening kits (Hampton Research, CA). Based on these screens, platelike crystals were found to grow at 17 °C within 24 h in 0.1 M calcium acetate, 0.05 M cacodylate, pH 6.5, 9% PEG 8000, 10% glycerol, and 12 mg/ml DBP–actin. These crystals grew to dimensions of 0:4 0:4 0:05 mm over a period of several days. The crystals were flash frozen in a nitrogen stream at 100 K and X-ray diffraction data were collected at the Brookhaven National Synchrotron Light Source, beamline X8C.
Results and discussion Intense research has been carried out during the past several years to establish structure-functional basis of certain functions of DBP. For example, our group and
that of the late Dr. Haddad employed affinity/photoaffinity labeling technique to identify Domain I (of DBP) as the specific area for vitamin D sterol-binding [21,26,28]. Domain I contains Trp145, the only Trp residue in human DBP. We reported earlier that oxidation of this residue with N-bromosuccinimide led to a complete loss of 25-OH-D3 -binding, supporting the important role of Trp145 and Domain I towards vitamin D sterol-binding [32]. In the present study we substantiated the above chemical mutation data with site-specific mutation of Trp145 to Ser, and determined 25-OH-D3 -binding property of this mutant by ligand-binding assays. Traditionally vitamin D sterol (25-OH-D3 and other metabolites and synthetic analogs)-binding property of DBP is ascertained by competitive radioligand-binding assay of this protein. Previous studies have demonstrated that the dissociation constant between DBP and 25-OH-D3 (and several of its metabolites and synthetic analogs) is in the nanomolar range, suggesting a very slow dissociation of the holo-protein [33–35]. Therefore, in a competitive radioligand binding assay samples of DBP are incubated with a fixed quantity of 3 H-25-OHD3 , which is competed out from the binding pocket by various concentrations of vitamin D sterols or their synthetic analogs [20–27,33,34]. Specific activity of 3 H25-OH-D3 , used in these assays, is usually very high (Ci/ mmol range) to reduce non-specific binding and increase the sensitivity of competition. Furthermore, in a companion assay a large excess of 25-OH-D3 is added to compete out majority of 3 H-25-OH-D3 from the ligandbinding pocket. Radioactivity, associated with the latter DBP samples (non-specific binding), is subtracted from earlier mentioned samples to obtain ‘specific binding’. A similar procedure is also used in assaying 25-OH-D3 in clinical samples [33,34]. We have recently shown that 25-OH-D3 -binding characteristic of recombinant human DBP (Gc2 isomorph) is very similar to that of native DBP [31]. Therefore, we surmised that competitive binding assays could be used to determine 25-OH-D3 -binding property of the recombinant Trp145-Ser mutant. We carried out such an assay of native DBP and Trp145-Ser mutant to observe that the mutant had a low 25-OH-D3 -binding capability (results not shown). Therefore, we performed a 3 H-25-OH-D3 -binding-analysis, in which specific 3 H25-OH-D3 -binding of native DBP was compared with that of the mutant. As shown in Fig. 2, the mutant variety retained only approximately 20% of the specific 3 H-25-OH-D3 -binding of the wild-type. This result conclusively established the essential role of Trp145 towards vitamin sterol-binding, and further implicated Domain I as the area of DBP that is responsible for such activity. However, this information could not be extrapolated to ascertain whether other domains of DBP are also responsible for the binding of vitamin D ligands.
N. Swamy et al. / Archives of Biochemistry and Biophysics 402 (2002) 14–23
Fig. 2. Specific 3 H-25-OH-D3 -binding assays of reDBP and Trp145Ser mutant. Samples of wild-type and mutant DBP (1:0 lg each) were incubated with 3 H-25-OH-D3 (3000 cpm, 0.15 pmol, dissolved in 10 ll of EtOH) at 4 °C followed by removal of unbound 3 H-25-OH-D3 by charcoal treatment, centrifugation, and counting of radioactivity in the supernatants (total binding). Non-specific binding was determined by carrying out the same assays in the presence of an excess of 25-OH-D3 (0:3 lg, 0.75 nmol). Specific binding was calculated by subtracting non-specific binding from total binding in each case. Each assay was run in triplicate.
Fig. 3. SDS–PAGE analysis of various structural domains of DBP to show their homogeneity. Domain I: 1–203, Domain I + II: 1–330, Domain III: 375–427. Positions of standard molecular weight markers (MW) are shown on the right of each panel.
We expressed different structural domains of DBP in bacteria, and produced largely homogeneous samples of Domain I (1–203), Domain I + II (1–330), and Domain
19
III (375–427) as evidenced by SDS–PAGE-analysis (Fig. 3). As with the Trp145-Ser mutant, first we performed a specific 3 H-25-OH-D3 -binding assay of the domains with a single large dose of 25-OH-D3 to determine their capability to bind 3 H-25-OH-D3 in a specific manner, compared to full-length reDBP. These assays showed that Domain I (1–203), Domain I + II (1–330), and Domain III (375–427) retained approximately 88, 76.9, and 5.8%, respectively, of the specific 3 H-25-OH-D3 -binding by the full-length reDBP (100%) (results not shown). Therefore we performed competitive binding assays of native DBP, full-length reDBP, Domain I (1–203), and Domain I + II (1–330). As shown in Fig. 4, concentrations of 25-OH-D3 , displacing half-maximally bound 3 H-25-OH-D3 were 17.5, 11.4, 19, and 12.1 nM for native DBP, full-length reDBP, Domain I (1–203), and Domain I + II (1–330) respectively. Therefore, results of two assays (specific binding and competitive binding) in combination showed that Domain II, in addition to Domain I, might participate in vitamin D sterol-binding, while Domain III is largely not required for such binding activity. However, these results did not provide any conclusive proof whether Domain II is directly involved in 25-OHD3 -binding. It is probable that presence of Domain II stabilized Domain I by ‘correct’ conformational changes to promote 25-OH-D3 -binding. Other indirect effects also could not be ruled out. In our recent studies with affinity analogs of 25-OHD3 and 1,25ðOHÞ2 D3 containing the affinity probe at different parts of the parent vitamin D structure, we demonstrated that the vitamin D sterol-binding pocket of DBP is highly constrained, and does not allow any ‘tumbling’ of the ligand inside the binding pocket [26,27]. Results of the present studies further suggested that gross structural integrity of the protein is also required for specific ligand-binding. It is noteworthy that similar steric restrictions of the ligand-binding pocket are also observed in 1; 25ðOHÞ2 D3 -binding pocket of the vitamin D receptor (VDR), the nuclear receptor for 1; 25ðOHÞ2 D3 [36–38]. However, DBP and VDR are entirely different classes of proteins, and our findings suggested that these two proteins might have evolved from a primordial gene and separated during evolution. In 1992 Haddad et al. employed actin-affinity chromatography to study the binding of the proteolytic fragments of DBP and demonstrated that a C-terminal fragment of DBP (350–403), which spans most of Domain III, is capable of binding actin independently [28]. However, these studies could not determine any relationship between vitamin D and its metabolites and actin in terms of their binding by DBP. In an earlier study we coupled the 3-hydroxyl group of 25-OH-D3 to Sepharose through a 12 atom tether, and affinity-purified fully functional DBP from human serum [30]. We argued that if vitamin D sterol- and
20
N. Swamy et al. / Archives of Biochemistry and Biophysics 402 (2002) 14–23
Fig. 4. Competitive 3 H-25-OH-D3 -binding assays of native DBP, full-length reDBP, and its structural domains (Domain I and Domain I + II). Briefly, samples of native DBP, reDBP, and various domains (1.0 lg each) in Tris–HCl buffer, pH 8.3, were incubated with 3 H-25-OH-D3 (0.1 pmol) and various doses of 25-OH-D3 (0.92-62 pmol) for 20 h at 4 °C followed by incubation with dextran-coated charcoal (4 °C) for 15 min, centrifugation, and radioactive counting of the supernatants. In each case non-specific binding was determined by carrying out the same assay in the presence of an excess of 25-OH-D3 . Specific binding was calculated by subtracting non-specific binding from total binding. For each protein sample, percent of specific binding ðB=B0 100Þ is plotted against concentration of the competitor (25-OH-D3 ). Concentrations of 25-OH-D3 required to displace 50% of 3 H-25-OH-D3 bound to the protein samples are denoted on the Y-axis by an arrow.
Fig. 5. Cartoon depicting the formation of the complex between DBP in human plasma and actin in rabbit muscle, and binding of the complex to 25OH-D3 that is chemically attached to a solid matrix. The bound-protein could be removed from the affinity matrix by the natural ligand (25-OH-D3 ), and the purified DBP–actin complex could be obtained by removing bound-25-OH-D3 by hydroxylapatite chromatography. Please note that change in conformation of DBP upon binding to 25-OH-D3 (chemically bound to matrix and free) is represented by slight change of its shape and graphic pattern.
actin-binding by DBP are largely independent processes, we might be able to employ this matrix to affinity purify a DBP–actin complex from human plasma, as depicted in Fig. 5.
Preparation of DBP–actin complex is described in the Materials and methods section. The denaturing SDS–PAGE analysis of the protein fraction, eluted with 25-OH-D3 from the affinity matrix, showed that the
N. Swamy et al. / Archives of Biochemistry and Biophysics 402 (2002) 14–23
21
Fig. 6. SDS–PAGE and native-PAGE analyses of DBP, actin, and DBP–actin complex. Positions of the standard molecular weight markers are denoted on the right of the left panel.
protein fraction was homogeneous, and contained only DBP and actin (Fig. 6, left panel). On the other hand, native PAGE analysis revealed a single band moving slower than DBP or actin (Fig. 6, right panel), proving the compositional homogeneity of the DBP–actin complex. In essence, we generated DBP–actin complex in plasma, and purified the complex by using 25-OHD3 –Sepharose affinity chromatography. These results suggested that the binding of vitamin D sterols (by DBP) might not influence the binding of actin in a significant manner. The above observation was further supported by the 3 H-25-OH-D3 -binding assays of DBP and DBP–actin
complex. As shown in Fig. 7, incubation of DBP or DBP–actin complex with 3 H-25-OH-D3 (0.139 pmol) in the presence of increasing amounts of 25-OH-D3 (0.97– 250 pmol) resulted in a dose-dependent displacement of 3 H-25-OH-D3 from the 25-OH-D3 -binding pocket of DBP in both cases. Additionally, the competitive binding curves were almost identical indicating that 25-OHD3 -binding characteristics of DBP–actin complex are very similar to those of DBP alone. McLeod et al. [39] demonstrated that hDBP, bound to 25-OH-D3 and 1,25ðOHÞ2 D3 , bound actin in a manner that was indistinguishable from apo-hDBP. Therefore, this information, in combination with the
Fig. 7. Competitive 3 H-25-OH-D3 -binding assays of native DBP and DBP–actin complex. Briefly, samples (0.2 lg, 4.0 pmol each) of native DBP (affinity-purified from human plasma) or DBP–actin complex (from human plasma and rabbit muscle) in 50 mM Tris–HCl, 150 mM NaCl, 1.5 mM of EDTA, and 0.1% Triton X-100, pH 8.3 were incubated with 3 H-25-OH-D3 (0.139 pmol) for 16 h at 4 °C in the absence or presence of various amounts of 25-OH-D3 (0.97–250 pmol). Following the incubation, the solutions were treated with ice-cold dextran-coated charcoal for 15 min at 4 °C and centrifuged at 5000g at 4 °C. The supernatants were mixed with scintillation cocktail and counted for radioactivity.
22
N. Swamy et al. / Archives of Biochemistry and Biophysics 402 (2002) 14–23
results described in this communication strongly suggested that vitamin D sterol- and actin-binding by DBP might be independent processes to a large extent. However, it should be emphasized that these results do not conclusively prove total independence of these binding processes. We argued that proof of our above-mentioned structure-functional analysis could be best achieved by determining the three-dimensional structure of DBP– actin by X-ray crystallography, and studying its prop erty in the presence of 25-OH-D3 . Recently a 2.3 A crystal structure of DBP–25-OH-D3 complex was reported, which revealed that the vitamin D-binding site is in the N-terminal part of the protein [40–42]. This communication also reports significant difference in the overall structure, as well as local folding pattern between the structures of DBP and serum albumin, a homologous protein, limiting the scope of structure determination of DBP–actin complex by molecular replacement technique based on the crystal structure of G-actin [43,44]. Prior to attempting to crystallize DBP–actin complex it was important to determine its stability under various conditions. Susceptibility of a protein in the apo- and holo-forms towards trypsin treatment is often used to determine its conformational changes upon ligandbinding. As shown in Fig. 8, lane 1, trypsinization of DBP primarily resulted in a prominent 25 kDa band and several low molecular weight peptides moving with the dye front. The amino acid sequence analysis of this 25 kDa trypsin-resistant band revealed that it belonged to
Fig. 8. Tryptic digestion of DBP, actin, and DBP–actin. Samples of native DBP, actin, and DBP–actin complex (30 lg each) in 50 mM Tris–HCl buffer, pH 7.4 were incubated with trypsin (20:1 protein:trypsin) for 14 h at 25 °C followed by analysis with 12% SDS– PAGE. Lane 1: DBP + trypsin; Lane 2: DBP–actin complex + trypsin; Lane 3: actin + trypsin; Lane 4: DBP–actin complex; MW: standard molecular weight markers (positions are shown on the right).
Domain I with five (5) amino acids truncated in the Nterminus. This indicated that while Domains II and III are trypsin susceptible, Domain I is largely resistant to trypsin cleavage. In contrast, trypsinization of actin proteolyzed the protein completely resulting in several low molecular weight peptides (Fig. 8, lane 3). On the other hand, trypsin treatment of DBP–actin complex showed two prominent bands of 52 and 34 kDa and other lower molecular weight peptides (Fig. 8, lane 2). Amino acid sequencing of the 52 and 34 kDa bands revealed that the 52 kDa band was DBP (5 amino acids shorter in Nterminus), and 34 kDa originated from actin (starting with amino acid position 69). These results indicated that DBP, particularly its Domains II and III were completely protected from trypsin-cleavage upon actinbinding, suggesting that Domain II, in addition to Domain III, might be involved in actin-binding. Similarly, actin, except its N-terminus, was also protected from tryptic digestion when complexed with DBP. Apparently, the interaction between DBP and actin stabilized both proteins and masked their trypsin sensitive sites. We also observed that DBP–actin complex was stable for at least 60 days at 4 °C. Furthermore, the complex was stable in the presence of up to 4.0 M urea (there was only partial dissociation in 5.3 M urea) at 25 °C as determined by native PAGE (results not shown). Observed stability of DBP–actin complex provided the basis for its crystallization studies, which were carried out according to the method described in the Matrials and methods section. Diffraction data collected from crystals were processed with DENZO/SCALEPACK [45] and showed the space group to be primitive , orthorhombic with unit cell dimensions a ¼ 80:2 A , and c ¼ 159:6 A . A full data set was colb ¼ 87:3 A . Systematic absences show the space lected to 2.5 A group to be P21 21 21 and Vm calculations indicated a single molecule of the complex in the asymmetric unit (Vm ¼ 2:9). The availability of crystals of the DBP–actin complex set the stage for studying its interaction with 25-OH-D3 . But, we have so far been unable to grow crystals of actin– DBP in the presence of 25-OH-D3 . Soaking of preexisting crystals of actin–DBP in crystallization buffer containing various concentrations of 25-OH-D3 resulted in cracking of the crystal. We have demonstrated that DBP–actin bound 25-OH-D3 similar to apo-DBP (Fig. 7). Conversely, McLeod et al. showed that DBP–25-OHD3 holo-protein bound DBP in a manner indistinguishable from apo-DBP [39]. Therefore, crystal cracking might reflect a ligand-induced conformational change in the complex, disrupting crystal contacts. However, a conclusive proof of this phenomenon would require knowledge of the three-dimensional structure of the DBP–actin complex. Currently structure determination is being pursued in our laboratory by a combination of
N. Swamy et al. / Archives of Biochemistry and Biophysics 402 (2002) 14–23
molecular replacement methods and heavy metal derivatization. Molecular replacement has provided a solution for the position of the actin in the complex, but the phases are inadequate for tracing the chain of the DBP. Molecular replacement using models based on the known structure of albumin, a DBP homolog, has not provided a solution. We are therefore working to obtain additional phase information from heavy metal derivatives using either isomorphous replacement or multiwavelength anomalous diffraction methods. In conclusion, we have probed the vitamin D steroland actin-binding properties of DBP with various techniques to establish structural basis for the multipleligand-binding properties of DBP. We have shown for the first time that structural domains I and II of DBP are required for vitamin D sterol-binding. We have also shown that in solution vitamin D sterol- and actinbinding are largely independent processes, although the binding of 25-OH-D3 likely results in conformational changes in the complex.
Acknowledgments The work was supported by a grant from the National Institute of Digestive, Diabetes and Kidney Diseases (#DK 44337 to R.R.). Assistance from Dr. Jan Krzyszt of Blusztajn in the preparation of the manuscript is gratefully acknowledged. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13]
N.E. Cooke, J.G. Haddad, Endocr. Rev. 4 (1996) 125–128. R. Ray, Proc. Soc. Exp. Biol. Med. 212 (1996) 305–312. J.G. Haddad, J. Steroid Biochem. Mol. Biol. 53 (1995) 579–582. N. Yamamoto, S. Homma, Proc. Natl. Acad. Sci. 88 (1991) 8539– 8543. N. Yamamoto, S. Homma, I. Millman, J. Immunol. 147 (1991) 273–280. G.B. Schneider, K.A. Benis, N.W. Flay, R.A. Ireland, S.N. Popoff, Bone 16 (1995) 657–662. N. Swamy, S. Ghosh, G.B. Schneider, R. Ray, J. Cell Biochem. 81 (2001) 535–546. N. Yamammoto, V.R. Naraparaju, Cancer Res. 57 (1997) 2187– 2192. Y. Koga, V.R. Naraparaju, N. Yamamoto, Proc. Soc. Exp. Biol. Med. 220 (1999) 20–26. H. Van Baelen, R. Bouillon, P. DeMoor, J. Biol. Chem. 255 (1980) 2270–2272. J.G. Haddad, Arch. Biochem. Biophys. 213 (1982) 538–544. P.J. Goldschmidt-Clermont, M.H. Williams, R.M. Galbraith, Biochem. Biophys. Res. Commun. 146 (1987) 611–617. S.E. Lind, D.B. Smith, P.A. Janmey, T.P. Stossel, J. Clin. Invest. 78 (1986) 736–742.
23
[14] W. Lee, R.M. Galbraith, N. Engl. J. Med. 326 (1992) 1335– 1341. [15] F. Yang, J.I. Brune, S.L. Naylor, R.L. Cupples, K.H. Naberhaus, B.H. Bowman, Proc. Natl. Acad. Sci. USA 82 (1985) 7994–7998. [16] N.E. Cooke, E.V. David, J. Clin. Invest. 76 (1985) 2420–2424. [17] H.S. Lichenstein, D.E. Lyons, M.M. Wurfel, D.A. Johnson, M.D. McGinley, J.C. Leidli, D.B. Trollinger, J.P. Mayer, S.D. Wright, M.M.. Zukowski, J. Biol. Chem. 269 (1994) 18149–18154. [18] F. Schoentgen, M. Metz-Boutigue, J. Jolles, J. Constans, P. Jolles, Biochem. Biophys. Acta 871 (1986) 189–198. [19] R. Ray, S.A. Holick, N. Hanafin, M.F. Holick, Biochemistry 25 (1986) 4729–4733. [20] R. Ray, R. Bouillon, H.G. Baelen, M.F. Holick, Biochemistry 36 (1991) 4809–4813. [21] R. Ray, R. Bouillon, H.G. Van Baelen, M.F. Holick, Biochemistry 30 (1991) 7638–7642. [22] N. Swamy, R. Ray, Arch. Biochem. Biophys. 319 (1995) 504– 507. [23] N. Swamy, R. Ray, Arch. Biochem. Biophys. 333 (1996) 139–144. [24] J.K. Addo, R. Ray, Steroids 63 (1998) 218–223. [25] J.K. Addo, N. Swamy, R. Ray, Steroids 64 (1999) 273–282. [26] N. Swamy, J.K. Addo, M.R. Uskokovic, R. Ray, Arch. Biochem. Biophys. 373 (2000) 471–478. [27] N. Swamy, A. Dutta, R. Ray, Biochemistry 36 (1997) 7432–7436. [28] J.G. Haddad, Y.Z. Hu, M.A. Kowalski, C. Laramore, K. Ray, P. Robzyk, N.E. Cooke, Biochemistry 31 (1992) 7174–7181. [29] R. Ray, in: M.F. Holick (Ed.), Vitamin D: Physiology, Molecular Biology and Clinical Applications, Humana Press, New Jersey, 1999, pp. 147–162. [30] N. Swamy, A. Roy, R. Chang, M. Brisson, R. Ray, Protein Expr. Purif. 6 (1995) 185–188. [31] N. Swamy, S. Ghosh, R. Ray, Protein Expr. Purif. 10 (1997) 115– 122. [32] N. Swamy, M. Brisson, R. Ray, J. Biol Chem. 270 (1995) 2636– 2639. [33] J.G. Hadadad, Methods Enzymol. 67 (1980) 449–459. [34] T.C. Chen, A.K. Turner, M.F. Holick, J. Nutr. Biochem. 1 (1990) 315–319. [35] R. Bouillon, W.H. Okamura, A.W. Norman, Endocr. Rev. 16 (1995) 200–257. [36] N. Rochel, J.M. Wurtz, A. Mitschler, B. Klaholz, D. Moras, Mol. Cell 5 (2000) 173–179. [37] N. Swamy, W. Xu, N. Paz, J-C. Hsieh, M.R. Haussler, G.J. Maalouf, S.C. Mohr, R. Ray, Biochemistry 39 (2000) 12162– 12171. [38] S.C. Mohr, N. Swamy, W. Xu, R. Ray, Steroids 66 (2001) 189– 201. [39] J.F. McLeod, M.A. Kowalski, J.G. Haddad, J. Biol. Chem. 264 (1989) 1260–1267. [40] C. Verboven, A. Rabijns, M. De Maeyer, H. Van Baelen, R. Bouillon, C. De Ranter, Nature Struct. Biol. 2 (2002) 131–136. [41] N.J. Vogelaar, U. Lindberg, C.E. Schutt, J. Mol. Biol. 220 (1991) 545–547. [42] C.C. Verboven, H.L. De Bondt, C. De Ranter, R. Bouillon, H. Van Baelen, J. Steroid Biochem. Mol. Biol. 54 (1995) 11–14. [43] W. Kabsch, H.G. Mannherz, D. Suck, E.F. Pai, K.C. Holmes, Nature 347 (1990) 37–44. [44] L.R. Otterbein, P. Graceffa, R. Dominguez, Science 293 (2001) 616–618. [45] W. Minor, D. Tomchick, Z. Otwinowski, Struct. Fold Des. 8 (2000) R105–110.