Effect of Hydrogen Ion Concentration and Buffer Composition on Ligand Binding Characteristics and Polymerization of Cow s Milk Folate Binding Protein

Transcription

1 Bioscience Reports, Vol. 21, No. 6, December 2001 ( 2002) Effect of Hydrogen Ion Concentration and Buffer Composition on Ligand Binding Characteristics and Polymerization of Cow s Milk Folate Binding Protein Jan Holm 1,3 and Steen Ingemann Hansen 2 Receiûed June 7, 2001 The ligand binding and aggregation behavior of cow s milk folate binding protein depends on hydrogen ion concentration and buffer composition. At ph 5.0, the protein polymerizes in Tris-HCl subsequent to ligand binding. No polymerization occurs in acetate, and binding is markedly weaker in acetate or citrate buffers as compared to Tris-HCl. Polymerization of ligand-bound protein was far more pronounced at ph 7.4 as compared to ph 5.0 regardless of buffer composition. Binding affinity increased with decreasing concentration of protein both at ph 7.4 and 5.0. At ph 5.0 this effect seemed to level off at a protein concentration of 10 6 M which is fold higher than at ph 7.4. The data can be interpreted in terms of complex models for ligand binding systems polymerizing both in the absence or presence of ligand (ph 7.4) as well as only subsequent to ligand binding (ph 5.0). KEY WORDS: Cows milk folate binding protein; ligand-mediated polymerization; ligand binding at ph 5.0 and buffer composition INTRODUCTION The folate binding proteins (FBP) in bovine and human milk exhibit a high degree of amino acid sequence homology [1]. The N-terminal amino acid sequence of the cell membrane folate receptor (FR) in malignant epithelial cells is homologous to that of soluble human milk FBP [2]. Based upon this evidence it seems justified to assume that soluble FBP can be used as a prototype receptor protein in studies devoted to the function of cell membrane FR and the interaction between receptor and ligand at the molecular level. A method involving a combination of cation exchange and methotrexate affinity chromatography has made it feasible to perform a large scale purification of FBP from cow s whey powder [3]. The purified protein has been characterized with regard to primary and secondary structure, physicochemical nature and ligand binding characteristics [3 11]. The protein possessed a concentration-dependent pronounced 1 Department of Clinical Chemistry, Central Hospital Bornholm, Roenne, DK-3700, BCMM- JH@BORA-dk. 2 Department of Clinical Chemistry, Central Hospital, Hollerød, DK-3400, Denmark. 3 To whom correspondence should be addressed Plenum Publishing Corporation

2 746 Holm and Hansen polymerization tendency at near-neutral ph, enhanced upon folate binding [4, 8]. At acidic ph (ph 5.0) polymerization was weak and did only occur subsequent to ligand binding [4, 8]. Two peculiar ligand binding characteristics at near-neutral ph, upward convex Scatchard plots and a binding affinity that depended on the FBP concentration suggested an interrelationship between ligand binding and polymerization equilibrium [5,8]. A few studies conducted at extremely high FBP concentrations suggested similar binding characteristics at ph 5.0 in Tris-HCl buffer [11]. This sharply contrasts earlier studies which showed a diminished folate binding at acidic ph in acetate buffer [8]. This discordance prompted us to study and compare polymerization and binding characteristics over a wide range of FBP concentrations at ph 5.0 in different buffer types. MATERIALS AND METHODS The radiochemical [ 14 C] folate (pteroylglutamate) with a specific activity of 52.4 Ci mol, and [ 3 H] folate with a specific activity of and 5.0 Ci mmol were obtained from Amersham International Ltd., Amersham, U.K. FBP was purified from cow s whey powder according to a previously published method which involved a combination of cation exchange chromatography and affinity chromatography [3]. All FBP solutions were dialyzed against 0.2 M acetate buffer, ph 3.5 at 4 C to remove endogenous folate. Unless otherwise stated, equilibrium dialysis experiments were performed as described previously [5] for periods of 20 hr in 0.17 M Tris-HCl buffer (37 C, ph 5.0) with FBP in the internal (1000 µl) and radioligand in the external solution (200 ml). Due to a large volume of external solution, the concentration of radioligand was kept constant during the entire dialysis experiment. Equilibrium dialysis with the following buffer types were also performed, acetate, phosphate, citrate and formate (0.2 M and ph 5.0). Triton X-100 at a concentration of 1 g l was added to both sides of the dialysis membrane [5]. Radioactivity was measured as previously described [5]. Gel Filtration Polymerization as a function of the concentration of ligand-bound FBP was studied within a range of µm by use of a gel filtration technique [5]. FBP dissolved in one of the following buffers, 0.17 M Tris-HCl (ph 7.4), 0.2 M acetate (ph 7.0 and 5.0) and M phosphate with 0.08 M NaCl (ph 7.4) was exposed to an excess labeled folate for three hours, and then applied to a Bio-Gel P-300 (Biorad Laboratories, Richmond, California, U.S.A.) column (2.6B37 cm) equilibrated and eluted with the same buffers at 5 C (flow rate 5 ml hr 1 ). The column was calibrated as described [5]. The FBP in the effluent was monitored by radioactivity measurements [5].

3 Folate Binding Protein 747 RESULTS Binding Characteristics of Folate at ph 5.0 as a Function of the Concentration of FBP. Influence of Buffer Composition Figure 1 shows saturation curves for binding of folate to FBP at ph 5.0 in 0.17 M Tris-HCl buffer. Data were obtained over four decades of FBP concentrations (21 µm-1.7 nm) by dilution of a stock solution of purified FBP. Maximum binding of folate was proportional to the concentration of FBP. The concentration of folate required for half saturation of binding, S 0.5, decreased with increasing dilution of the FBP solution leading to a parallel increase in the overall affinity equal to the reciprocal value of S 0.5. However, the effect of FBP concentration on S 0.5 seemed to level off after a 100-fold dilution approaching a minimum value (maximum binding affinity) of 1 nm. Binding studies at ph 5.0 were also performed with other buffer types. No qualitative differences estimated from the appearance of the Scatchard plot (downward concavity indicative of apparent positive cooperativity) were observed at 20 µm FBP after substitution of Tris-HCl with acetate buffer (Fig. 2A). At 100-fold lower concentrations of FBP (0.2 µm) there was a drastic difference between binding characteristics in Tris-HCl and acetate (Fig. 2B) or citrate (data not shown) buffers where binding affinity and maximum binding were diminished. Substitution of Tris- HCl by phosphate or formate buffers at low FBP concentrations had no effect on folate binding (data not shown). Fig. 1. Saturation curves for high-affinity binding of radiolabeled folate at ph 5.0 to solutions of FBP obtained by dilution of a stock solution (maximum folate binding, 21 µm), undiluted ( ); diluted 10- times ( ), 100 times ( ), 1000 times ( ), and 10,000 times ( ). Equilibrium dialysis experiments in 0.17 M Tris-HCl buffer, ph 5.0 at 37 C.

4 748 Holm and Hansen Fig. 2. Scatchard plots of high-affinity folate binding at ph 5.0. Ordinate, bound free radioligand; abscissa, bound radioligand. A, high (20 µm) FBP concentrations in 0.17 M Tris-HCl buffer ( ) and 0.2 M aceteate buffer ( ). B, low (100 times dilution, 0.2 µm) FBP concentrations in 0.17 M Tris-HCl buffer and 0.2 M acetate buffer (symbols, cf. above). Polymerization as a Function of the Concentration of Ligand-bound FBP at ph 7.4 and ph 5.0. Gel filtration profiles were obtained with decreasing concentrations of radioligand-bound FBP at ph7 4 in Tris-HCl (Fig. 3A, upper part), phosphate (Fig. 3A, lower part) and acetate, ph 7.0 (Fig. 3B, upper part). In all three cases polymerization decreased with declining concentrations of FBP converging towards the monomeric state at low concentrations (1 nm) of FBP. Only a slight polymerization of FBP occurred at ph 5.0 in acetate buffer (Fig. 3B, lower part).

5 Folate Binding Protein 749 Fig. 3. Degree of polymerization P (n-mer) as a function of the concentration of FBP. The apparent molecular size of the FBP peak, decreasing with increasing elution volume, was determined by gel filtration with FBP solutions saturated with radiolabeled folate. A, Gel filtration at ph 7.4 in Tris-HCl buffer (upper part) and phosphate buffer (lower part). B, Gel filtration at ph 7.0 (upper part) and ph 5.0 (lower part) in acetate buffer. Samples were applied to a Bio-Gel P-300 column calibrated, equilibrated and eluted with the appropriate buffers. The concentration of FBP (µm), 25.6 ( ), 12.8 ( ), 1.28 ( ), 0.12 ( ), ( ) and ( ).

6 750 Holm and Hansen Relationship between Polymerization of Ligand-bound FBP and the Affinity for Ligand Binding Figure 4A shows a plot of S 0.5 (cf. Fig. 1) versus the FBP concentration at ph 5.0 in Tris-HCl buffer. A similar curve obtained at ph 7.4 in Tris-HCl buffer is inserted (data taken from reference 5). Both curves take a similar parallel course within the FBP concentration range M where S 0.5 exhibits a decrease proportional to the decrease in the FBP concentration. At ph 5.0 only a slight and declining decrease in S 0.5 occurred at FBP concentration below 10 6 M meaning that the impact of the FBP concentration on S 0.5 seemed to level off with binding affinity approaching a maximum value (cf. Fig. 1). At ph 7.4 (Fig. 4A) the effect of the Fig. 4. A, Ligand concentration at half saturation of binding (S 0.5 )at 37 C in 0.17 M Tris-HCl at ph 7.4 (dotted line, data from reference 5) and ph 5.0 ( ) vs. the FBP concentration. B, Polymerization vs. the FBP concentration. At ph 7.4 in Tris-HCl ( ) and phosphate ( ), at ph 7.0 in acetate ( ); at ph 5.0 in acetate ( ); and Tris-HCl (dotted line data from reference 11).

7 Folate Binding Protein 751 FBP concentration on S 0.5 did also seem to level off, but at a 100-fold lower FBP concentration where the binding affinity approached a maximum value 100-fold greater than that at ph 5.0. Polymerization of ligand-bound FBP is very pronounced at ph 7.4 but diminishes with decreasing concentrations of FBP converging towards a state of equilibrium between monomeric and dimeric forms at M FBP (Fig. 4B) where dependence of S 0.5 on the concentration of FBP begins to level off. At ph 5.0 polymerization of ligand-bound FBP in Tris-HCl buffer, 3-mer (data from reference, 11), far less pronounced than at ph 7.4, seemed to converge towards a monomeric state at FBP concentrations of M (Fig. 4B) where dependence of S 0.5 on the FBP concentration exhibits a sharp decrease (Fig. 4A). Polymerization of ligand-bound FBP at ph 5.0 was far less pronounced in acetate buffer where only monomeric forms were present at FBP concentrations below 10 6 M (Fig. 4B). DISCUSSION Cow s milk FBP consists of 222 amino acid residues; its molecular weight is 30 kda based on the amino acid composition and carbohydrate content [6]. Multiple isoelectric points within the ph range 7 8, probably due to variations in the carbohydrate content, means that the protein is electrically neutral at near-neutral ph [3,10]. The unliganded protein exhibits a remarkable polymerization tendency at near-neutral ph, whereas only the monomeric form seems to be present at ph 5 [4,8]. Polymerized FBP at ph 8 is far more hydrophobic and stable against thermal unfolding than monomeric FBP at ph 5 [9,10]. Ligand binding induced profound changes in the secondary structure of FBP as evidenced by an altered CD-spectrum, quenching of tryptophan fluorescence, increased stability against thermal as well as guanidinium chloride-induced unfolding of FBP [7, 9]. The changes were accompanied by a two-fold enhancement of polymerization upon ligand binding, both at ph 7.4 and 5.0 [4,8]. The polymerized FBP was far less hydrophobic in the liganded state, and there was no difference between water solubility of liganded FBP at ph 7.4 and 5.0 [10]. In general, the physicochemical differences between unliganded FBP at ph 7.4 and 5.0 seemed to disappear subsequent to ligand binding. The conformational changes in the FBP molecule resulted in a very tight irreversible binding of ligand only dissociable at low acidic ph, and did moreover seem to protect the ligand against degradation [11]. The ligand binding characteristics of FBP are complex. A number of theoretical models describing binding of small molecules to reversibly associating macromolecules as well as ligand-induced polymerization phenomena have been published [12 14]. Non-linear Scatchard plots and ligand binding affinities depending on the concentration of the binding macromolecule indicates competitive interaction between a ligand and a reversibly polymerizing protein. Two mechanisms have been proposed: A ligand-facilitated one where ligand binding enhances polymerization of polymeric protein and a ligand-mediated one where ligand binding is obligatory for polymerization of monomeric protein [12 14]. According to the present as well as previous data [5,8,11] the former system operates at ph 7.4, whereas the latter system

8 752 Holm and Hansen seems to be active at ph 5.0. A concentration-dependent equilibrium between polymeric and monomeric FBP at ph 7.4 means predominance of the monomer in highly diluted solutions of FBP. Assuming that the ligand only binds tightly to the monomer the binding affinity will approach its maximum value at low FBP concentrations. Ligand binding will by diminishing the concentration of unliganded monomer favor deaggregation of unliganded polymer making more monomers available for binding, which results in convex upward Scatchard plots similar to those typically seen in positive cooperativity; the plots become linear at low concentrations of (monomer) FBP [12 14]. The formation of stable liganded polymers of FBP will by removing liganded monomer from the reacting system give rise to a further enhancement of ligand binding. At ph 5.0 FBP is a monomeric, cationic and hydrophilic protein with a low folding stability. Ligand binding leads to a concentration-dependent polymerization of the protein accompanied by changes in conformational structure and folding stability. The binding affinity at ph 5.0 in highly diluted solutions of (monomeric) FBP (Figs. 1 and 4A) seems to approach a maximum value 100-fold lower than at ph 7.4. Ligand-mediated polymerization at ph 5.0 depends on ligand type being far more pronounced in the presence of folate as compared to 5-methyltetrahydrofolate which possesses a weaker binding affinity for FBP than folate [12], and buffer type, being less pronounced in acetate buffer as compared to Tris buffer (Fig. 4B) where binding affinity for folate is higher (Fig. 2B). Hence, these two observations lend support to the idea of a close association between the ligand binding process and the polymerization phenomena. The impact of acetic and citric acids on binding characteristics of FBP could imply a possible role of these metabolites in folate homeostasis in an acidic intracellular environment. Two types of ligand-mediated polymerization (ligand binding precedes polymerization) have been proposed [12 14]. The first one involves a concentrationdependent association between one liganded and one unliganded monomer. Dimerization inhibits ligand binding (reduces affinity) by removal of unligandedmonomer from the reacting system meaning that very high concentrations of ligand are required to reverse dimerization via binding to released unliganded monomer. Scatchard plots describing this process are downward convex, i.e. similar to negative cooperativity, and becomes linear at low (monomeric) protein concentrations where binding affinity approaches its maximum value. The increase in binding affinity with decreasing concentrations of FBP (Fig. 1 and Fig. 4A) over a wide concentration range [11] is consistent with the proposed model, but the upward convex Scatchard plots (Fig. 2, and reference 11) and the high polymers, 3-mer (Fig. 4B) observed at high FBP concentrations [11] are more compatible with the second model implying a concentration-dependent association between two liganded monomers. The removal of liganded monomers from the reacting system after dimerization will enhance additional ligand binding and give rise to upward convex Scatchard plots. However, at low FBP concentrations (Fig. 1) where dimers coexisted with monomers (Fig. 4B), and there seemed to be a shift from upward to downward convex Scatchard plots (data not shown) the observations were consistent with the first model. To summarize, a concentration-dependent competition between the two types of ligand-mediated polymerization could be an attractive possibility. The second type

9 Folate Binding Protein 753 being most active at high FBP concentrations in Tris-HCl buffer where liganded monomers tend to associate with each other forming dimers and even higher more stable polymers. The first type being more predominant at lower FBP concentrations, particularly in acetate buffer, where liganded monomers prefer association with unliganded monomers to form dimers thereby reducing the binding affinity of the ligand. ACKNOWLEDGMENTS We appreciate the valuable technical assistance of Jytte Rasmussen. REFERENCES 1. Svendsen, I., Hansen, S. I., Holm, J., and Lyngbye, J. (1982) Carlsberg Res. Commun. 47: Campbell, I. G., Jones, T. A., Foulkes, W. D., and Trowsdale, J. (1991) Cancer Res. 51: Svendsen, I., Martin, B., Pedersen, T. G., Hansen, S. I., Holm, J., and Lyngbye, J. (1979) Carlsberg Res. Commun. 44: Pedersen, T. G., Svendsen, I., Hansen, S. I., Holm, J., and Lyngbye, J. (1980) Carlsberg Res. Commun. 45: Hansen, S. I., Holm, J., Lyngbye, J., Pedersen, T. G., and Svendsen, I. (1983) Arch. Biochem. Biophys. 226: Svendsen, I., Hansen, S. I., Holm, J., and Lyngbye, J. (1984) Carlsberg Res. Commun. 49: Kaarsholm, K. C., Kolstrup, A.-M., Danielsen, S. E., Holm, J., and Hansen, S. I. (1993) Biochem. J. 292: Salter, D. N., Scott, K. J., Slade, H., and Andrews, P. (1981) Biochem. J. 193: Sigurskjold, B. W., Christensen, T., Hansen, S. I., and Holm, J. (1997) in: Chemistry and Biology of Pteridines and Folates, (W. Pfleiderer and H. Rokos, eds.) Blackwell Science, pp Holm, J. and Hansen, S. I. (2001). Biosci. Rep. 21:(in press). 11. Holm, J. and Hansen, S. I. (2001). Biosci. Rep. 21:(in press). 12. Nichol, L. W., Jackson, W. J. H., and Winzor, D. J. (1967) Biochemistry 6: Nichol, L. W. and Winzor, D. J. (1976) Biochemistry 15: Cann, J. R. (1978) Methods Enzymol. 48: