Get 20M+ Full-Text Papers For Less Than $1.50/day. Subscribe now for You or Your Team.

Learn More →

Physicochemical Characterization of Casein Phosphopeptide-Amorphous Calcium Phosphate Nanocomplexes *

Physicochemical Characterization of Casein Phosphopeptide-Amorphous Calcium Phosphate... THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 280, No. 15, Issue of April 15, pp. 15362–15369, 2005 © 2005 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A. Physicochemical Characterization of Casein Phosphopeptide- Amorphous Calcium Phosphate Nanocomplexes* Received for publication, December 1, 2004, and in revised form, January 14, 2005 Published, JBC Papers in Press, January 17, 2005, DOI 10.1074/jbc.M413504200 Keith J. Cross, N. Laila Huq, Joseph E. Palamara, John W. Perich, and Eric C. Reynolds‡ From the Centre for Oral Health Science, School of Dental Science, The University of Melbourne, Melbourne, Victoria 3010, Australia Many techniques have been used to investigate the ultra- Milk caseins stabilize calcium and phosphate ions and make them available to the neonate. Tryptic digestion of structure of the casein micelles. Although the structural details the caseins yields phosphopeptides from their polar N- are still being elucidated, the casein micelles are believed to be terminal regions that contain clusters of phosphoryl- roughly spherical particles with a radius of 100 nm, dispersed ated seryl residues. These phosphoseryl clusters have in a continuous phase of water, salt, lactose, and whey proteins been hypothesized to be responsible for the interaction (4). The calcium phosphate isolated after exhaustive hydrazine between the caseins and calcium phosphate that lead to deproteination of micelles has been reported to exhibit a fine the formation of casein micelles. The casein phos- and uniform granularity under the electron microscope with phopeptides stabilize calcium and phosphate ions the particles consisting of small subunits of 2.5-nm diameter (5, through the formation of complexes. The calcium phos- 6). The calcium phosphate, present as nanometer-sized ion phate in these complexes is biologically available for clusters, and caseins are not covalently bound; hence the casein intestinal absorption and remineralization of subsur- micelle is known as an association colloid (7). Nevertheless, the face lesions in tooth enamel. We have studied the struc- casein micelles are extremely stable and can withstand boiling, ture of the complexes formed by the casein phosphopep- freeze-drying, and the addition of salt and ethanol. It is be- tides with calcium phosphate using a range of lieved that the amphipathic, glycosylated C-terminal end of physicochemical techniques including x-ray powder dif- -casein protrudes from the micelle surface forming a so-called fraction, scanning electron microscopy, transmission “hairy layer” that sterically stabilizes the complexes (8). The electron microscopy, and equilibrium binding analyses. literature on casein interactions has been reviewed by Horne The amorphous nature of the calcium phosphate phase (9), and a model of the casein micelle has been formulated that was confirmed by two independent methods: x-ray pow- accounts for many of the physicochemical properties of the der diffraction and selected area diffraction. In solution, micelle. The model involves electrostatic interactions between the ion activity product of a basic amorphous calcium phosphate phase was the only ion product that was a colloidal calcium phosphate particles and multiple - and -ca- function of bound phosphate independent of pH, con- sein molecules and hydrophobic interactions between the -, -, sistent with basic amorphous calcium phosphate being and -caseins forming a cross-linked network (9). Electron mi- the phase stabilized by the casein phosphopeptides. De- croscopy of casein micelles (10) has provided evidence that the tailed investigations of calcium and calcium phosphate caseins are organized into tubular structures within the binding using a library of synthetic homologues and micelle. analogues of the casein phosphopeptides have revealed The casein micelles serve as a carrier of calcium phosphate that although the fully phosphorylated seryl-cluster mo- providing the neonate with a bioavailable source of calcium and tif is pivotal for the interaction with calcium and phos- phosphate ions for bone and teeth formation (3). It has been phate, other factors are also important. In particular, postulated that the ability of casein to form stable complexes calcium binding and calcium phosphate stabilization by with calcium phosphate is intrinsic to a general mechanism for the peptides was influenced by peptide net charge, avoiding pathological calcification and regulating calcium flow length, and sequence. in tissues and biological fluids containing high concentrations of calcium (11). The ability of casein micelles to maintain calcium and phos- Bovine milk contains 30 mM calcium and 22 mM inorganic phate ions in a soluble and bioavailable state is retained by the phosphate in solution with most of the calcium (68%) and tryptic multiphosphorylated peptides of the caseins known as phosphate (47%) associated with the proteins  -,  -, -, and S1 S2 the casein phosphopeptides (CPP) (12). The major tryptic CPP -casein in casein micelles (1, 2). The  -,  -, and -caseins S1 S2 are -CN(1–25) (sequence 1 below) and  -CN(59 –79) (se- S1 have a number of Ser(P) residues in a specific motif, Ser(P) - quence 2 below) with smaller amounts of  -CN(46 –70) (se- S2 Glu , that is involved in the interaction with calcium quence 3 below) and  -CN(1–21) (sequence 4 below) (13, 14). S2 phosphate (3). These peptides all contain the cluster sequence motif Ser(P) - Glu with three contiguous phosphoserines. This peptide motif * This work was supported by National Health and Medical Research is thought to be critical for calcium and calcium phosphate Council Grant IO 209042 and by the Dairy Research and Development binding by these peptides (12). The sequences of the four major Corporation. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. The abbreviations used are: CPP, casein phosphopeptide(s); ACP, ‡ To whom correspondence should be addressed: Centre for Oral amorphous calcium phosphate; CN, casein; DCPD, dicalcium phosphate Health Science, School of Dental Science, The University of Melbourne, dihydrate; HA, hydroxyapatite; OCP, octacalcium phosphate; SEM, Victoria 3010, Australia. Tel.: 61-3-9341-0270; Fax: 61-3-9341-0236; scanning electron microscopy; TEM, transmission electron microscopy; E-mail: e.reynolds@unimelb.edu.au. HPLC, high performance liquid chromatography. 15362 This paper is available on line at http://www.jbc.org This is an Open Access article under the CC BY license. Casein Phosphopeptide-Amorphous Calcium Phosphate 15363 the use of Boc-Ser(PO Ph )-OH in the Boc mode of peptide synthesis casein tryptic phosphopeptides are shown with the motif un- 3 2 followed by platinum-catalyzed hydrogenolytic deprotection of the pro- derlined: sequence 1 (-CN(1–25)), Arg -Glu-Leu-Glu-Glu-Leu- tected Ser(PO Ph )-containing peptides (14, 28, 29). The synthetic pep- 3 2 Asn-Val-Pro-Gly-Glu-Ile-Val-Glu-Ser(P)-Leu-Ser(P) -Glu -Ser- 3 2 tides were capped with an acyl group at the N terminus and a methyl- 25 59 Ile-Thr-Arg ; sequence 2 ( -CN(59 –79)), Gln -Met-Glu-Ala- S1 amine group at the C terminus. In the case of Ac-Ile-Val-Pro-Asn- Glu-Ser(P)-Ile-Ser(P) -Glu -Ile-Val-Pro-Asn-Ser(P)-Val-Glu-Gln- 3 2 Ser(P)-Val-Glu-Glu-NHMe, a homologue of  -CN(71–78), a glutamyl S1 79 46 Lys ; sequence 3 ( -CN(46 –70)), Asn -Ala-Asn-Glu-Glu-Glu- residue was substituted for Gln to avoid a problematic synthesis S2 associated with a C-terminal glutamine residue. The preparations of Tyr-Ser-Ile-Gly-Ser(P) -Glu -Ser(P)-Ala-Glu-Val-Ala-Thr-Glu- 3 2 70 1 the other synthetic analogues have been described previously (29 –31). Glu-Val-Lys ; and sequence 4 ( -CN(1–21)), Lys -Asn-Thr- S2 The peptides were purified by reversed phase HPLC, and the purity was Met-Glu-His-Val-Ser(P) -Glu -Ser-Ile-Ile-Ser(P)-Gln-Glu-Thr- 3 2 confirmed by capillary electrophoresis, amino acid composition, se- Tyr-Lys . quence analysis, mass spectrometry, and NMR spectroscopy (13, 14, 23, The CPP stabilize calcium and phosphate ions under neutral 28 –31). and alkaline conditions forming metastable solutions that are Calcium Binding to  -CN(59–79), -CN(1–25),  -CN(63–70), and S1 S1 -CN(71–78)—All calcium binding experiments were performed using supersaturated with respect to the basic calcium phosphate S1 a modification of the ultrafiltration method of Marsh (32) without the phases (15). Under these conditions, the CPP bind their equiv- addition of phosphate. The solutions were buffered using 100 mM Tris- alent weights of calcium and phosphate (16). The CPP are HCl to a pH value of 7.0, 7.5, 8.0, 8.5, or 9.0. Ionic strength was adjusted formed in vivo by normal digestion of casein and, because they to 0.16 – 0.19 using NaCl. The reaction mixtures were incubated at room are relatively resistant to further proteolytic degradation, ac- temperature for 18 h. Subsequently, less than 10% of the total volume cumulate in the distal portion of the small intestine (17–21). It was collected as ultrafiltrate by centrifugation at 1000  g for 15 min using Centrifree MPS-1 Micropartition cells (Amicon) equipped with has been proposed that this accumulation together with the YM-3 (3000 molecular weight exclusion limit) or YM-1 (1000 molecular ability of the peptides to form soluble complexes with calcium weight exclusion limit) membranes. These membranes were demon- phosphate are responsible for the enhanced intestinal calcium strated not to retain calcium or phosphate ions or ion pairs. Calcium ion absorption that has been observed even in vitamin D-deficient concentrations of acidified samples were measured at 422.7 nm by animals consuming dietary CPP (17–21). In addition, CPP in- atomic absorption spectroscopy using PerkinElmer instrument model crease the calcification of in vitro cultured embryonic rat bone, 373, with the addition of 1% LaCl to prevent phosphate interference. Calcium concentrations ranged from 0.5 to 21.5 mM for  -CN(59 –79) and again the mechanism is suggested to be associated with the S1 and -CN(1–25), 0.3–5.5 mM Ca for the  -CN(63–70) peptide, and S1 ability of the peptide to form soluble complexes with calcium 0.05–1.5 mM Ca for the  -CN(71–78) peptide. The calcium concen- S1 and phosphate ions (22). Furthermore, CPP-calcium phosphate trations of the original reaction mixture after centrifugation to demon- complexes have been shown to be anticariogenic and to remi- strate no precipitation and of the ultrafiltrate were determined, and the neralize early stages of enamel caries in animal and human peptide-bound calcium was calculated as the difference between these studies (12, 15, 23–25). In summary, the ability to stabilize values. Calcium binding by the peptides was modeled by assuming the number of independent ion-binding sites/peptide ( ), where each site calcium phosphate and thereby enhance mineral solubility and Ca has a dissociation constant (K ) given by the following equation. bioavailability (26) confers upon the CPP the potential to be biological delivery vehicles for calcium and phosphate (27). Peptide][Ca ] As part of our long term investigation into the structure- K  (Eq. 1) [PeptideCa ] function relationships of proteins involved in biomineralization and calcium phosphate stabilization, we have studied the in- The dissociation constants for the peptide/calcium complexes (K ) and the number of calcium ion-binding sites/mol of peptide ( ) were de- teraction of tryptic phosphopeptides from milk caseins with the Ca termined by a nonlinear least squares fit to the equation, amorphous and crystalline phases of calcium phosphate. In this paper, we report our investigations of the calcium and [Ca ] Ca free phosphate binding properties of the two major CPP, -CN(1– [Ca ]  (Eq. 2) bound K  Ca ] free 25) and  -CN(59 –79). We demonstrate that the experimen- S1 tally determined ion activity product of a basic amorphous where K  K . Calcium Phosphate Binding to  -CN(59–79) and -CN(1–25) Ho- calcium phosphate (ACP) phase best correlates with the cal- S1 mologues and Analogues—Binding was initiated by adding phosphate cium bound by the peptide  -CN(59 –79) over a range of S1 to solutions containing 14 mM calcium and 4 mg/ml peptide buffered calcium and phosphate concentrations and sample pH ranging using 100 mM Tris-HCl to a pH value of 7.0, 7.5, 8.0, 8.5, or 9.0. Sodium from 7.0 to 9.0. Furthermore, we delineate the regions and chloride was added to bring the ionic strength for each sample to 0.16. residues of these peptides that are responsible for calcium For the homologues and analogues, binding was determined at pH 7.0 phosphate stabilization. We report the effect of peptide length, or 9.0. The reaction mixtures were incubated and ultrafiltered as de- scribed above for calcium binding. The phosphate concentration ranged residue type, and order of acidic residues on the binding to from0to9mM. Phosphate concentration was determined colorimetri- calcium and calcium phosphate using a library of synthetic cally (33), with absorbance measured at 660 nm on a PerkinElmer 552 analogues. Finally, we describe the ultrastructure of the casein Spectrophotometer. Calcium and phosphate concentrations in the orig- phosphopeptide-calcium phosphate nanocomplex as deter- inal solution and the ultrafiltrate were determined. The peptide bound mined using a range of physicochemical techniques including calcium and peptide bound phosphate were taken as the difference powder diffraction x-ray crystallography, scanning electron mi- between the total and free calcium and phosphate, respectively. To confirm that no precipitation had occurred during incubation, the sam- croscopy (SEM), and transmission electron microscopy (TEM). ples were centrifuged at 17,000  g for 5 min, and the calcium and EXPERIMENTAL PROCEDURES phosphate concentrations were determined prior to ultrafiltration. Ul- Preparation of Casein Phosphopeptides—The casein phosphopep- trafiltrates were examined spectroscopically at 214 nm and were found tides -CN(1–25) and  -CN(59 –79) were selectively precipitated from to contain no peptide. Peptide-bound calcium phosphate is expressed as S1 a tryptic digest of casein using calcium chloride and ethanol and further the number of calcium ions ([Ca ] ) and phosphate ions ([P ] )/ bound i bound purified by anion exchange fast protein liquid chromatography and mol of peptide. The ion activity products for various phases of calcium reversed phase HPLC (13). The purity of the peptides was assessed by phosphate were determined from the free calcium and phosphate con- matrix-assisted laser desorption ionization time-of-flight mass spec- centrations and pH using an iterative computational procedure that trometry, capillary electrophoresis, amino acid composition, and se- calculates the ion activity coefficients using the expanded Debye- quence analyses (13, 14). Prior to sequence analysis, the labile phos- Hu ¨ ckel equation. This procedure takes into account ion pairs CaHPO , phoseryl residues were converted to S-ethyl cysteinyl residues by CaH PO , and CaPO ; the dissociation of H PO and H O; and the 2 4 4 3 4 2 -elimination (13). ionic strength (32, 34). The activity of the CaOH ion was explicitly Preparation of Synthetic Peptides—The peptide Ac-Glu-Ser(P)-Ile- assumed to be negligible. Dissociation constants at 37 °C were used Ser(P) -Glu -NHMe corresponding to  -CN(63–70) was prepared by from the following sources: H PO (35); HPO (36); CaH PO and 3 2 S1 2 4 4 2 4 15364 Casein Phosphopeptide-Amorphous Calcium Phosphate CaHPO (37); and CaPO (38). The sources of the solubility products at 4 4 37 °C were hydroxyapatite (HA) (39), octacalcium phosphate (OCP) (40), dicalcium phosphate dihydrate (DCPD) (37), and ACP (41). Cal- cium and phosphate binding was modeled as occurring at multiple independent sites within the peptide with each site having a K de- d2 fined by the following equation. Peptide][Ca ][P ] K  (Eq. 3) d2 [PeptideCa P ] The number of phosphate binding sites was determined by nonlinear least squares fits to an equation of the form of Equation 2, in which bound and free phosphate were the variables. The value of K from these fits can be shown to be approximately given by the equation, d2 K  (Eq. 4) where K is the calcium ion binding constant for the peptide. The number of calcium binding sites was similarly determined using the equation, Ca i free 2 2 Ca    Ca  (Eq. 5) bound initial K  P i free FIG.1. Calcium ion binding by casein phosphopeptides. Titra- 2 tion of peptides with calcium ions. Bound calcium is plotted as a func- where [Ca ] is the amount of calcium bound with no added Initial tion of the free calcium ion concentration. Units of bound calcium are phosphate. calcium ions bound/mol of peptide. (f)  -CN(59 –79), () -CN(1–25), S1 Preparation of CPP-Calcium Phosphate—CPP was dissolved at 10 and (Œ) Ac-Glu-Ser(P)-Ile-Ser(P)-Ser(P)-Ser(P)-Glu-Glu-NHMe, the g/liter in Milli-Q water. 1.6 M CaCl and 1 M Na HPO were added 2 2 4 synthetic octapeptide corresponding to  -CN(63–70). The solid curves S1 slowly by syringe pump (0.5–1.0 ml/min) to the CPP solution in a are nonlinear least squares fits of the data to Equation 2. pH-stat held at pH 9.0 by automatic titration of 5 N NaOH. After 60 min of titration the final calcium and inorganic phosphate concentration TABLE I were 100 and 60 mM, respectively. The colloidal CPP-calcium phosphate Calcium binding characteristics of the casein phosphopeptides nanocomplexes were then concentrated 5-fold by microfiltration Peptide pH V K Ca d through a 0.1-m Sartorius filter (polyolefin) in a mini-Sarticon micro- b b filtration system and washed with 3 volumes of Milli-Q water to remove -Casein(63–70) 8.00 3.52  0.13 2.71  0.18 S1 free calcium and phosphate ions. The final CPP-calcium phosphate -Casein(59–79) 8.00 6.84  0.45 5.07  0.61 S1 solutions were then lyophilized. -Casein(1–25) 8.00 4.64  0.17 2.28  0.25 Preparation of -CN(1–25)-Calcium Phosphate—The -CN(1–25) v  number of calcium-binding sites/molecule of peptide. Ca phosphopeptide was dissolved at 1.67 g/liter in Milli-Q water. To the Means  S.D. (mM). peptide solution, 50 mM CaCl and 21.3 mM Na HPO were added 2 2 4 stepwise slowly. The pH was held at 9.0 by automatic titration of 50 mM Calcium Phosphate Binding to  -CN(59–79) and -CN(1– NaOH. The final -CN(1–25)-ACP solution was then lyophilized and S1 25)—Fig. 2 shows the titration of  -CN(59 –79) with inor- then redissolved in H O. S1 Powder Diffraction of CPP-ACP—The powdered samples were ganic phosphate in the presence of 14 mM calcium and at mounted in an aluminum holder and were placed in a Sintag pad V various pH values. Peptide-bound calcium and phosphate were x-ray diffractometer, operating with a Cu K (  1.5406 Å) x-ray determined by the difference in the total and free calcium and source at 45 kV and 40 mA. The x-ray diffraction data were collected phosphate concentrations, respectively. The quantity with a germanium solid state collector and scanned from 10° to 100° in 2 2 0 2 0 [Ca ]  [Ca ] , where [Ca ] is the calcium 2 at 2° min (step size 0.03°). Phase identification was accomplished bound bound bound bound in the absence of phosphate, represents the additional by comparison of the x-ray diffraction data pattern with the JCPDS powder diffraction file, produced by the International Center for Dif- calcium bound in the presence of phosphate. This quantity has fraction Data. been plotted versus free phosphate (P ) in Fig. 2A. Fig. 2B SEM and TEM of CPP-ACP—The sample for TEM was prepared shows a plot of bound phosphate ([P ] ) versus free P . i bound i from powdered material that was crushed and placed on a 3-mm TEM Nonlinear least squares fits were performed to determine the grid. The TEM grids were covered with a Butvar/Formvar (ProSciTec, number of mol of calcium and phosphate bound and the effec- Thuringowa Central, Australia) film to support the powder. Transmis- tive dissociation constant K. The fits to Equation 5 showed a sion electron micrographs were taken on a 200-kV JEOL 200CX instru- 2 0 ment operating at 200 kV. strong correlation between the value of [Ca ] and the bound 2 0 For SEM, specimens were sprinkled on an aluminum stub covered value of K.If[Ca ] was not constrained, it adopted un- bound with adhesive conductive carbon tabs and gold sputter-coated (Gold reasonable values, and the value of K determined from fits to Sputter coater S150B, Edwards, UK). Observations were made with a the phosphate data disagreed with those determined from the field emission SEM instrument (Philips XL 30 FEG, Eindhoven, The 2 0 calcium data. Fixing [Ca ] at the experimental value of Netherlands) operating at 20 kV using the secondary electron mode. bound the bound calcium ion concentration at zero phosphate gave RESULTS values of K in reasonable agreement with those derived from Calcium Binding to  -CN(59–79), -CN(1–25),  -CN(63– the phosphate data. The number of calcium ion-binding sites/ S1 S1 70), and  -CN(71–78)—Fig. 1 shows the calcium-binding mol of peptide ( ) was not sensitive to the value of S1 Ca 2 0 curves for (a)  -CN(59 –79) and (b)  -CN(63–70) with the [Ca ] . The values reported in Table II are appropriately S1 S1 bound bound calcium/mol of peptide ([Ca ] ) being plotted as a weighted averages of the experimental parameters determined bound function of free calcium ion concentration. Table I summarizes from repeated measurements. Fig. 2C, a plot of [Ca ] bound the values of the dissociation constants and the number of versus [P ] for the samples at different pH, reveals a linear i bound calcium ion-binding sites/peptide from fitting the data for  - relationship between bound calcium and bound phosphate. S1 CN(59 –79), -CN(1–25), and  -CN(63–70) to Equation 2 us- This demonstrates that the binding of calcium ions and phos- S1 ing a nonlinear least squares algorithm. In the case of the phate ions is co-dependent. The number of calcium phosphate C-terminal region  -CN(71–78) homologue containing only a binding sites/mol of peptide shows a strong correlation with the S1 single phosphoserine, no calcium binding was observed. length of the peptide as shown in Fig. 5. Furthermore, the ratio Casein Phosphopeptide-Amorphous Calcium Phosphate 15365 FIG.2. Calcium phosphate binding by  -CN(59–79). Titration of  -CN(59 –79) with inorganic phosphate in the presence of 14 mM S1 S1 calcium at various pH values is shown. A, bound calcium versus free phosphate. B, bound inorganic phosphate versus free P . C, bound calcium versus bound inorganic phosphate. of [Ca ] versus [Pi] increases as the peptides become z  0.00 to z  0.20. At each value of z, the ion activity product bound bound shorter (Table II). was plotted as a function of the number of calcium-binding Determination of the Calcium Phosphate Phase Stabilized by sites, and a quartic polynomial was fitted to the data. The -CN(59–79)—To determine the calcium phosphate phase correlation coefficients from these fits were then plotted S1 bound by  -CN(59 –79), the binding of calcium phosphate against z, and a quadratic polynomial fitted through these data S1 was related to the ion activity products of the various biolog- points to determine the optimal value of z. Fig. 4 shows the ically relevant calcium phosphate phases. The bound calci- variation in the correlation coefficient as a function of z for the um/mol of peptide at various pH values was plotted as a amorphous phases. function of the ion activity products of the following biologi- Calcium Phosphate Binding to Homologues and Ana- 2 2 cally relevant phases: DCPD (Ca )(HPO ) (Fig. 3A) (42), logues—In addition to -CN(1–25) and  -CN(59 –79), the cal- 4 S1 2  3 2 3 OCP (Ca ) (H )(PO ) (Fig. 3B), HA (Ca ) (PO ) (OH ) cium phosphate binding characteristics of a series of synthetic 4 4 3 5 4 3 2 3 (Fig. 3C), and a basic ACP (Ca ) (PO ) (OH ) , where homologues and analogues were investigated as shown in Ta- 3z 4 2 2z z  0.0877 (Fig. 3D). Only the ACP phase produced a function ble II. The aim of this investigation was to delineate the resi- that was independent of pH (Fig. 3D), indicating that this was dues important for calcium phosphate stabilization and to in- the phase stabilized by  -CN(59 –79). Two types of ACP were vestigate the importance of phosphorylation, peptide length, S1 considered, acidic ACP phases represented by (Ca ) - residue type, and order. The synthetic peptides, Ile-Val-Pro- 3z/2 3 2 (PO ) (HPO ) and basic ACP phases represented by Asn-Ser(P)-Glu-Glu and Gln-Met-Glu-Ala-Glu, corresponding 4 2z 4 z 2 3 (Ca ) (PO ) (OH ) . These formulae are capable of rep- to the C and N termini of  -CN(59 –79), respectively, did not 3z 4 2 2z S1 resenting a wide range of calcium phosphate compositions bind calcium or stabilize calcium phosphate. The importance of ranging from HA (z  ⁄3) in the basic range, to tricalcium the multiphosphorylated motif Glu-Ser(P)-Ile/Leu-Ser(P) -Glu 3 2 phosphate (z  0), OCP (z  ⁄3) and DCPD (z  2) in the acidic was further investigated using a series of analogues. A plot of range. To determine the optimum stoichiometry, ion activity  versus the number of residues in the peptide is shown in Ca products were calculated at a number of points in the interval Fig. 5, which produced a linear correlation coefficient of 0.9956. 15366 Casein Phosphopeptide-Amorphous Calcium Phosphate TABLE II Calcium and phosphate binding characteristics of CPP and analogues a b c Peptide pH Charge   K Calcium:P Ca Pi d i -Casein(59–79) 7.0 13.52 16.86  0.35 11.22  0.56 1.63  0.17 1.50  0.11 S1 7.5 13.85 23.2  1.6 15.8  1.1 0.74  0.10 1.46  0.20 8.0 13.99 21.73  0.68 14.26  0.38 0.257  0.021 1.524  0.088 8.5 14.10 22.3  2.3 14.3  1.6 0.064  0.020 1.56  0.33 9.0 14.29 21.43  0.39 14.05  0.79 0.072  0.016 1.52  0.11 -Casein(59–79) N74D 7.0 14.20 15.3  1.5 10.32  0.56 1.48  0.22 S1 9.0 15.27 22.35  0.33 14.25  0.57 1.569  0.086 -Casein(63–70) 9.0 11.00 9.64  0.26 5.60  0.14 0.132  0.015 1.721  0.089 S1 E I EE 9.0 10.99 7.85  0.22 5.11  0.15 1.538  0.090 -Casein(59–63) 9.0 2.00 Precipitated S1 -Casein(71–78) 9.0 2.99 Precipitated S1 -Casein(1–25) 7.0 12.61 15.65  0.16 10.55  0.98 1.48  0.15 9.0 13.26 24.76  0.63 16.67  0.47 1.485  0.079 -Casein(14–21) 9.0 11.00 12.01  0.30 6.35  0.16 0.084  0.024 1.891  0.095 Ser(P) -Glu 9.0 8.00 8.76  0.19 4.804  0.083 1.824  0.071 3 2 Thr(P) -Glu 9.0 7.99 4.90  0.10 2.69  0.16 1.82  0.16 3 2 Ser-Thr(P) -Glu 9.0 6.00 1.861  0.051 1.104  0.061 1.69  0.14 2 2 Ser-Ser(P) -Glu 9.0 6.00 1.711  0.084 1.16  0.12 1.47  0.22 2 2 Glu-Ser(P) -Glu 9.0 7.00 2.045  0.029 1.10  0.11 1.86  0.22 2 2 Asp-Ser(P) -Glu 9.0 7.00 1.917  0.088 0.992  0.052 1.93  0.19 2 2 (Ala-Ser(P) -Ala-Glu 9.0 9.00 4.948  0.049 2.899  0.065 1.670  0.037 (Ala-Ser(P)) -Ala-Glu 9.0 7.00 Precipitated (Ala-Ser(P)) -Ala-Glu 9.0 5.00 Precipitated Ala-Glu-Ala 9.0 4.00 Precipitated Ala-Glu-Ala 9.0 3.00 Precipitated Ala-Glu 9.0 3.00 Precipitated Charges calculated using the pK parameters listed in Bundi and Withrich (46) and Bienkiewicz and Lumb (47). b 2 number Ca -binding sites/mol of peptide. Ca number of P -binding sites/mol of peptide. Pi i Mean  S.D. (mM). E  Glu, I  Ile,  Thr(P). FIG.3. pH dependence of ion activ- ity products of calcium phosphate phases and peptide-bound calcium. The number of mol of bound calcium at various pH values plotted as a function of the activity product of the DCPD 2 2 phase (Ca )(HPO )(A), the OCP phase 2 3 (Ca ) (PO ) (H )(B), the HA phase 4 4 3 2 3 (Ca ) (PO ) (OH )(C), and the basic 5 4 3 2 3 ACP phase (Ca ) (PO ) (OH ) 3.0877 4 2 0.1754 (D). The curves in A–C are interpolated curves using the method of Stineman (42) provided as an aid to visualization and to emphasize that a single function does not relate the bound calcium to the particular ion activity product. The curve in D is a quartic polynomial fit to all the data points. Casein Phosphopeptide-Amorphous Calcium Phosphate 15367 FIG.6. Powder diffraction spectra of CPP-ACP complexes. Shown are the X-ray powder diffraction spectra of CPP-ACP (A) and heat-treated CPP-ACP (B) showing peaks characteristic of an apatitic phase. The vertical bars show the expected positions of peaks in the x-ray powder spectrum of hydroxyapatite. phosphate bound by the CPP and their analogues increased as the pH increased. The calcium phosphate phase that was found to bind to  -CN(59 –79) was a basic ACP phase of the com- S1 position, Ca (PO ) (OH) . 3.0877 4 2 0.1754 X-ray Powder Diffraction of CPP-ACP—A number of samples FIG.4. Correlation of bound calcium and ion activity products of CPP-ACP and -CN(1–25)-ACP, were studied using x-ray pow- as a function of calcium phosphate phase. Plot showing the vari- ation in correlation coefficient as a function of phase composition (z) for der diffraction. An understanding of the degree of the crystallin- quartic polynomial fits to the number of calcium ions bound/mol of ity of the CPP-ACP complex can be obtained from the line broad- peptide as a function of the ion activity products of the acidic (dashed ening of the peak intensity spectrum for the polycrystalline line) and basic ACP (solid line) phases. powder. A sample spectrum is shown in Fig. 6A. The absence of sharp features in these spectra was consistent with an amor- phous calcium phosphate phase associated with the CPP and the individual peptides. A few peaks in the spectrum of the -CN(1– 25)-ACP complex were observed and assigned to traces of crys- talline NaCl and KCl contaminating the sample. Heat treatment of CPP-ACP produces material with less bio-available calcium and phosphate, and x-ray powder diffraction patterns indicated some partial crystallinity with conversion to a disordered apatitic phase upon heat treatment, as shown in Fig. 6B. SEM and TEM Characterization of CPP-ACP—Neither SEM nor TEM showed any crystallites in the CPP-ACP samples. Selected area electron diffraction over the particles also did not show any ring pattern characteristic of crystal structure. The SEM and TEM images of CPP-ACP and the selected area diffraction pattern therefore confirm the amorphous nature of the complex (Fig. 7). DISCUSSION FIG.5. Correlation of bound calcium to peptide length in cal- cium phosphate-peptide complexes. Plot showing correlation be- Calcium Binding—As shown in Table I,  -CN(59 –79) max- S1 tween the number of residues in the peptide and the number of calcium imally bound 6.84  0.45 mol of Ca /mol of peptide, whereas ion-binding sites/mol of peptide for the CPP-ACP complexes and their 2 -CN(1–25) bound 4.64  0.17 mol of Ca /mol of peptide. The analogues. The linear fit has a correlation coefficient of 0.993. shorter peptide  -CN(63–70) bound 3.52  0.13 mol of Ca / S1 mol of peptide. All of the measurements were performed at pH The substitution of any of the Ser(P) residues by Ser, Glu, or 8. The number of calcium ion-binding sites ( ) appears to be Ca Asp residues in the peptide Ser(P) -Glu or the substitution of 3 2 determined by charge for the longer peptides, the amount of Thr(P) residues by Thr in the Thr(P) -Glu peptide substantially calcium bound being sufficient to nearly cancel the intrinsic reduced the peptides ability to stabilize calcium phosphate. negative charge of the peptides. Further, the substitution of Ser(P) by Thr(P) was also associ- The charge on the peptide  -CN(63–70) is only slightly S1 ated with a reduced ability to stabilize calcium phosphate. lower than that on the other two peptides, and yet the amount Three contiguous phosphoseryl residues were necessary for full of calcium bound is significantly lower than would be expected activity, because the (Ala-Ser(P)) peptide, although containing based on charge neutralization. A structural explanation ap- four noncontiguous phosphoseryl residues, poorly stabilized pears to be the most plausible for the markedly lower than calcium phosphate. expected calcium binding ability of the shorter peptide. Inspec- In summary, in the presence of phosphate ions, the CPP and tion of calcium-binding proteins whose structures have been their analogues bound additional calcium (over and above that deposited in the Protein Data Bank suggests that calcium ions bound in the absence of phosphate). The linear dependence of the excess calcium bound versus the bound phosphate, inde- pendent of pH, implies that a specific phase of calcium phos- K. J. Cross, N. Laila Huq, J. E. Palamara, and E. C. Reynolds, phate was being bound (Fig. 2C). The amount of calcium and unpublished observations. 15368 Casein Phosphopeptide-Amorphous Calcium Phosphate calcium phosphate phase. Once the peptide is bound to the ACP phase, weaker interactions allow the remaining polar residues to interact with the ACP core particle, thus accounting for the correlation between peptide length and the number of mol of calcium and phosphate bound by these peptides. Given that the entire length of the peptide interacts with the calcium phos- phate phase and the stringent steric requirements associated with the formation of calcium ion-binding sites, it is unlikely that peptides separately bind calcium ions that are not associ- ated with the calcium phosphate phase. Characterization of the Bound Calcium Phosphate Phase— Fig. 4 shows plots of the calcium bound by the peptide -CN(1– 25) versus ion activity products for the four relevant calcium phosphate phases: HA, DCPD, OCP, and a basic amorphous calcium phosphate having the composition Ca (PO ) (OH) 3z 4 2 2z where z  0.0877. The calcium phosphate phase stabilized by -CN(1–25) is identified by the functional dependence of bound calcium (or phosphate) on the ion activity product independent of sample pH. Inspection of Fig. 4 shows that the amount of calcium bound is not a pH-independent function of the ion activity products for the phases HA, DCPD, or OCP. Fig. 4 shows that the bound calcium best correlates with a basic amorphous calcium phosphate phase Ca (PO ) (OH) , with 3z 4 2 2z z  0.0877  0.0022. The stabilization of a basic calcium phosphate phase is consistent with the observation that these complexes form over a pH range from 5.0 to 9.0. The ratio of  : was found to increase as the length of the Ca Pi peptide decreased (Table II). One possible explanation for this FIG.7. Electron microscopy of CPP-ACP complexes. a, SEM observation would be that the shorter peptides bind both a image of CPP-ACP; the segments of the scale bar represent a distance of 10 m. b, TEM image of CPP-ACP at 100,000. c, selected area calcium phosphate phase and independent calcium ions. How- diffraction pattern of CPP-ACP. Note the absence of evidence for crys- ever, as discussed above, the stringent requirements associated talline material. with the formation of calcium ion-binding sites makes this an unlikely explanation. usually have between six and eight oxygen atoms in their first These separate observations can be rationalized by propos- coordination sphere. The structure of the calcium-binding sites ing a model of the ACP core that consists of two calcium is based on distorted octahedral geometries with “small bite” phosphate phases: a calcium-poor phase with a Ca:P ratio of ligands, such as carboxylate groups, contributing two oxygen 1.5, such as Ca (PO ) , forming the core of the ACP particles as 3 4 2 atoms at some vertices of the octahedra. Water molecules usu- suggested by Meyer and Eanes (45), and a calcium-rich phase ally account for only one or two oxygen atoms/binding site. with a Ca:P ratio of 2.0, such as Ca (PO )(OH), that is in 2 4 Apart from interactions with side chain oxygen atoms, calcium contact with the peptide. As the peptides become shorter, a ions also interact with oxygen atoms from the backbone car- given number of peptides are able to fully cover a smaller bonyl moieties. The structural constraints implied by the mul- surface area; hence the size of the complex formed decreases as tiple interactions between the shorter peptide and the calcium the peptides become shorter. The thickness of the ions that are required to balance the peptide charge are likely Ca (PO )(OH) phase interacting with peptide is proposed to 2 4 to be impossible to satisfy, resulting in the lower calcium bind- remain constant as the complexes shrink in size, whereas the ing ability of the shorter peptide. The structural explanation is size of the relatively more soluble Ca (PO ) core decreases in 3 4 2 consistent with our previous reports of calcium ion-dependent, size, resulting in a steady increase in the ratio of  : as the Ca Pi structural features in the H NMR spectra of  -CN(59 –79) S1 peptides decrease in length. It is interesting that substitution and -CN(1–25) complexed with calcium (43, 44). We have of Ser(P) with Thr(P) reduced calcium phosphate stabilization shown that these peptides adopt structures consisting of loops activity. This together with possible steric hindrance in phos- and turns and that the specific structure adopted depends on phorylation of Thr residues may help explain why to date all of the peptide sequence outside of the calcium-binding motif the proteins known to stabilize calcium phosphate contain clus- Ser(P) -Glu . 3 2 ters of Ser(P) residues. Calcium Phosphate Binding—In the presence of phosphate The ion product that best correlates with the total bound ions, the CPP and their analogues bound additional calcium calcium by  -CN(59 –79) in a pH-independent manner is that (over and above that bound in the absence of phosphate). The S1 of a basic amorphous calcium phosphate phase having the linear dependence of the excess calcium bound versus the approximate composition Ca (PO ) (OH) . Given that bound phosphate, independent of pH, implies that a specific 3.0877 4 2 0.1754 each  -CN(59 –79) maximally binds 14 phosphate ions, the phase of calcium phosphate was being bound (Fig. 2C). The S1 composition of the  -CN(59 –79)-ACP complex is approxi- amount of calcium and phosphate bound by the CPP and their S1 mately ( -CN(59 –79)-(ACP) ) , where n is a small integer. analogues increased as the pH increased. S1 7 n Extensive studies using a library of synthetic homologues The number of calcium-binding sites in the calcium phos- phate complexes ( ) correlates strongly with the number of and analogues have revealed that the Ser(P) -Glu motif found Ca 3 2 in most of the CPP is a specific sequence of acidic residues that aminoacyl residues in the peptide used to stabilize the CPP- ACP complex. This is consistent with a model of the CPP-ACP has a strong affinity for calcium phosphate. However, the con- formational preferences of the intact major peptides enable complexes in which the phosphorylated sequence motif, Ser(P)- Ser(P)-Ser(P)-Glu-Glu, is required to initialize binding to the stabilization of the maximum amount of ACP. Casein Phosphopeptide-Amorphous Calcium Phosphate 15369 206 –211 Acknowledgment—We thank Fiona Webber for technical assistance. 25. Shen, P., Cai, F., Nowicki, A., Vincent, J., and Reynolds, E. C. (2001) J. Dent. REFERENCES Res. 80, 2066 –2070 26. Kitts, D. D., Yuan, Y. V., Nagasawa, T., and Moriyama, Y. (1992) Brit. J. Nutr. 1. Walstra, P., and Jenness, R. (1984) Dairy Chemistry and Physics, John Wiley 68, 765–781 & Sons, Inc., New York 27. Ferraretto, A., Signorile, A., Gravaghi, C., Fiorilli, A., and Tettamanti, G. 2. Van Hooydonk, A. C. M., Boerrigter, I. J., and Hagedoorn, H. G. (1986) Neth. (2001) J. Nutr. 131, 1655–1661 Milk Dairy J. 40, 297–313 28. Perich, J. W., Johns, R. B., and Reynolds, E. C. (1992) Aust. J. Chem. 45, 3. Holt, C., and Sawyer, L. (1988) Protein Eng. 2, 251–259 385–394 4. Schmidt, D. G. (1982) Dev. Dairy Chem. 1, 61– 86 29. Perich, J. W., Kelly, D. P., and Reynolds, E. C. (1992) Int. J. Pept. Prot. Res. 40, 5. McGann, T. C., Kearney, R. D., Buchheim, W., Posner, A. S., Betts, F., and 81– 88 Blumenthal, N. C. (1983) Calcif. Tissue Int. 35, 821– 833 30. Perich, J. W. (1991) Method Enzymol. 201, 225–233 6. McGann, T. C., Buchheim, W., Kearney, R. D., and Richardson, T. (1983) 31. Perich, J. W., Terzi, E., Carnazzi, E., Seyer, R., and Trifilieff, E. (1994) Int. J. Biochim. Biophys. Acta 760, 415– 420 Pept. Prot. Res. 44, 305–312 7. de Kruif, C. G. (1999) Int. Dairy J. 9, 183–188 32. Marsh, M. E. (1989) Biochemistry 28, 346 –352 8. Holt, C., and Horne, D. S. (1996) Neth. Milk Dairy J. 50, 85–111 33. Itaya, K., and Ui, M. (1966) Clin. Chim. Acta 14, 361–366 9. Horne, D. (1998) Int. Dairy J. 8, 171–177 34. Smales, F. C. (1972) Calcif. Tissue Res. 8, 304 –319 10. Dalgleish, D. G., Spagnuolo, P. A., and Goff, H. D. (2004) Int. Dairy J. 14, 35. Bates, R., and Acree, S. (1943) J. Res. Natl. Bureau Stand. 30, 129 1025–1031 36. Bjerrum, N. (1929) Kgl. Danske. Videnskab. Selskab. Math. fys. Medd. 9, 5–206 11. Holt, C., Wahlgren, N. M., and Drakenberg, T. (1996) Biochem. J. 314, 37. Gregory, T. M., Moreno, E. C., Brown, W. E. (1970) J. Res. Natl. Inst. Stand. 1035–1039 74A, 461 12. Reynolds, E. C., Black, C. L., Cai, F., Cross, K. J., Eakins, D., Huq, N. L., 38. Chughtai, A., Marshall, R., and Nancollas, G. H. (1968) J. Phys. Chem. 72, Morgan, M. V., Nowicki, A., Perich, J. W., Riley, P. F., Shen, P., Talbo, G., 208 –211 and Webber, F. (1999) J. Clin. Dent. 10, 86–88 39. McDowell, H., Gregory, T. M., and Brown, W. E. (1977) J. Res. Natl. Bur. 13. Reynolds, E. C., Riley, P. F., and Adamson, N. J. (1994) Anal. Biochem. 217, Stand. Sect. A. 81A, 273–281 277–284 40. Shyu, L. J. (1982) in The solid/solution interface-a kinetic study of the crystal- 14. Adamson, N. J., Riley, P. F., and Reynolds, E. C. (1993) J. Chromatogr. 646, lization of calcium fluoride and phosphate. Ph.D. thesis, State University of 391–396 New York, Buffalo, NY 15. Reynolds, E. C., Cain, C. J., Webber, F. L., Black, C. L., Riley, P. F., Johnson, I. H., and Perich, J. W. (1995) J. Dent. Res. 74, 1272–1279 41. Nancollas, G. H., LoRe, M., Perez, L., Richardson, C., and Zawacki, S. J. (1989) Anat. Rec. 224, 234 –241 16. Reeves, R. E. (1958) Science 128, 472 17. Mykkanen, H. M., and Wasserman, R. H. (1980) J. Nutr. 110, 2141–2148 42. Stineman, R. W. A. (1980) Creative Computing 6, 54 –57 43. Huq, N. L., Cross, K. J., and Reynolds, E. C. (1995) Biochim. Biophys. Acta 18. Lee, Y. S., Noguchi, T., and Naito, H. (1983) Brit. J. Nutr. 49, 67–76 19. Lee, S. L., and Veis, A. (1980) J. Pept. Prot. Res. 16, 231–232 1247, 201–208 20. Meisel, H., and Fristar, H. (1988) Biol. Chem. Hoppe-Seyler. 369, 1275–1279 44. Cross, K. J., Huq, N. L., Bicknell, W., and Reynolds, E. C. (2001) Biochem. J. 21. Sato, R., Noguchi, T., and Naito, H. (1986) J. Nutr. Sci. Vitaminol. 32, 67–76 356, 277–285 22. Gerber, H. W., and Jost, R. (1986) Calcif. Tissue Int. 38, 350 –357 45. Meyer, J. L., and Eanes, E. D. (1978) Calcif. Tissue Res. 25, 59–68 23. Reynolds, E. C. (May 14, 1991) U. S. Patent 5,015,628 46. Bundi, A., and Wu ¨ thrich, K. (1979) Biopolymers 18, 285–297 24. Reynolds, E. C., Cai, F., Shen, P., and Walker, G. D. (2003) J. Dent. Res. 82, 47. Bienkiewicz, E. A., and Lumb, K. J. (1999) J. Biomolec. NMR 15, 203–206 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Biological Chemistry American Society for Biochemistry and Molecular Biology

Physicochemical Characterization of Casein Phosphopeptide-Amorphous Calcium Phosphate Nanocomplexes *

Loading next page...
 
/lp/american-society-for-biochemistry-and-molecular-biology/physicochemical-characterization-of-casein-phosphopeptide-amorphous-Pm0ClEcNiH

References (40)

Publisher
American Society for Biochemistry and Molecular Biology
Copyright
Copyright © 2005 Elsevier Inc.
ISSN
0021-9258
eISSN
1083-351X
DOI
10.1074/jbc.m413504200
Publisher site
See Article on Publisher Site

Abstract

THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 280, No. 15, Issue of April 15, pp. 15362–15369, 2005 © 2005 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A. Physicochemical Characterization of Casein Phosphopeptide- Amorphous Calcium Phosphate Nanocomplexes* Received for publication, December 1, 2004, and in revised form, January 14, 2005 Published, JBC Papers in Press, January 17, 2005, DOI 10.1074/jbc.M413504200 Keith J. Cross, N. Laila Huq, Joseph E. Palamara, John W. Perich, and Eric C. Reynolds‡ From the Centre for Oral Health Science, School of Dental Science, The University of Melbourne, Melbourne, Victoria 3010, Australia Many techniques have been used to investigate the ultra- Milk caseins stabilize calcium and phosphate ions and make them available to the neonate. Tryptic digestion of structure of the casein micelles. Although the structural details the caseins yields phosphopeptides from their polar N- are still being elucidated, the casein micelles are believed to be terminal regions that contain clusters of phosphoryl- roughly spherical particles with a radius of 100 nm, dispersed ated seryl residues. These phosphoseryl clusters have in a continuous phase of water, salt, lactose, and whey proteins been hypothesized to be responsible for the interaction (4). The calcium phosphate isolated after exhaustive hydrazine between the caseins and calcium phosphate that lead to deproteination of micelles has been reported to exhibit a fine the formation of casein micelles. The casein phos- and uniform granularity under the electron microscope with phopeptides stabilize calcium and phosphate ions the particles consisting of small subunits of 2.5-nm diameter (5, through the formation of complexes. The calcium phos- 6). The calcium phosphate, present as nanometer-sized ion phate in these complexes is biologically available for clusters, and caseins are not covalently bound; hence the casein intestinal absorption and remineralization of subsur- micelle is known as an association colloid (7). Nevertheless, the face lesions in tooth enamel. We have studied the struc- casein micelles are extremely stable and can withstand boiling, ture of the complexes formed by the casein phosphopep- freeze-drying, and the addition of salt and ethanol. It is be- tides with calcium phosphate using a range of lieved that the amphipathic, glycosylated C-terminal end of physicochemical techniques including x-ray powder dif- -casein protrudes from the micelle surface forming a so-called fraction, scanning electron microscopy, transmission “hairy layer” that sterically stabilizes the complexes (8). The electron microscopy, and equilibrium binding analyses. literature on casein interactions has been reviewed by Horne The amorphous nature of the calcium phosphate phase (9), and a model of the casein micelle has been formulated that was confirmed by two independent methods: x-ray pow- accounts for many of the physicochemical properties of the der diffraction and selected area diffraction. In solution, micelle. The model involves electrostatic interactions between the ion activity product of a basic amorphous calcium phosphate phase was the only ion product that was a colloidal calcium phosphate particles and multiple - and -ca- function of bound phosphate independent of pH, con- sein molecules and hydrophobic interactions between the -, -, sistent with basic amorphous calcium phosphate being and -caseins forming a cross-linked network (9). Electron mi- the phase stabilized by the casein phosphopeptides. De- croscopy of casein micelles (10) has provided evidence that the tailed investigations of calcium and calcium phosphate caseins are organized into tubular structures within the binding using a library of synthetic homologues and micelle. analogues of the casein phosphopeptides have revealed The casein micelles serve as a carrier of calcium phosphate that although the fully phosphorylated seryl-cluster mo- providing the neonate with a bioavailable source of calcium and tif is pivotal for the interaction with calcium and phos- phosphate ions for bone and teeth formation (3). It has been phate, other factors are also important. In particular, postulated that the ability of casein to form stable complexes calcium binding and calcium phosphate stabilization by with calcium phosphate is intrinsic to a general mechanism for the peptides was influenced by peptide net charge, avoiding pathological calcification and regulating calcium flow length, and sequence. in tissues and biological fluids containing high concentrations of calcium (11). The ability of casein micelles to maintain calcium and phos- Bovine milk contains 30 mM calcium and 22 mM inorganic phate ions in a soluble and bioavailable state is retained by the phosphate in solution with most of the calcium (68%) and tryptic multiphosphorylated peptides of the caseins known as phosphate (47%) associated with the proteins  -,  -, -, and S1 S2 the casein phosphopeptides (CPP) (12). The major tryptic CPP -casein in casein micelles (1, 2). The  -,  -, and -caseins S1 S2 are -CN(1–25) (sequence 1 below) and  -CN(59 –79) (se- S1 have a number of Ser(P) residues in a specific motif, Ser(P) - quence 2 below) with smaller amounts of  -CN(46 –70) (se- S2 Glu , that is involved in the interaction with calcium quence 3 below) and  -CN(1–21) (sequence 4 below) (13, 14). S2 phosphate (3). These peptides all contain the cluster sequence motif Ser(P) - Glu with three contiguous phosphoserines. This peptide motif * This work was supported by National Health and Medical Research is thought to be critical for calcium and calcium phosphate Council Grant IO 209042 and by the Dairy Research and Development binding by these peptides (12). The sequences of the four major Corporation. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. The abbreviations used are: CPP, casein phosphopeptide(s); ACP, ‡ To whom correspondence should be addressed: Centre for Oral amorphous calcium phosphate; CN, casein; DCPD, dicalcium phosphate Health Science, School of Dental Science, The University of Melbourne, dihydrate; HA, hydroxyapatite; OCP, octacalcium phosphate; SEM, Victoria 3010, Australia. Tel.: 61-3-9341-0270; Fax: 61-3-9341-0236; scanning electron microscopy; TEM, transmission electron microscopy; E-mail: e.reynolds@unimelb.edu.au. HPLC, high performance liquid chromatography. 15362 This paper is available on line at http://www.jbc.org This is an Open Access article under the CC BY license. Casein Phosphopeptide-Amorphous Calcium Phosphate 15363 the use of Boc-Ser(PO Ph )-OH in the Boc mode of peptide synthesis casein tryptic phosphopeptides are shown with the motif un- 3 2 followed by platinum-catalyzed hydrogenolytic deprotection of the pro- derlined: sequence 1 (-CN(1–25)), Arg -Glu-Leu-Glu-Glu-Leu- tected Ser(PO Ph )-containing peptides (14, 28, 29). The synthetic pep- 3 2 Asn-Val-Pro-Gly-Glu-Ile-Val-Glu-Ser(P)-Leu-Ser(P) -Glu -Ser- 3 2 tides were capped with an acyl group at the N terminus and a methyl- 25 59 Ile-Thr-Arg ; sequence 2 ( -CN(59 –79)), Gln -Met-Glu-Ala- S1 amine group at the C terminus. In the case of Ac-Ile-Val-Pro-Asn- Glu-Ser(P)-Ile-Ser(P) -Glu -Ile-Val-Pro-Asn-Ser(P)-Val-Glu-Gln- 3 2 Ser(P)-Val-Glu-Glu-NHMe, a homologue of  -CN(71–78), a glutamyl S1 79 46 Lys ; sequence 3 ( -CN(46 –70)), Asn -Ala-Asn-Glu-Glu-Glu- residue was substituted for Gln to avoid a problematic synthesis S2 associated with a C-terminal glutamine residue. The preparations of Tyr-Ser-Ile-Gly-Ser(P) -Glu -Ser(P)-Ala-Glu-Val-Ala-Thr-Glu- 3 2 70 1 the other synthetic analogues have been described previously (29 –31). Glu-Val-Lys ; and sequence 4 ( -CN(1–21)), Lys -Asn-Thr- S2 The peptides were purified by reversed phase HPLC, and the purity was Met-Glu-His-Val-Ser(P) -Glu -Ser-Ile-Ile-Ser(P)-Gln-Glu-Thr- 3 2 confirmed by capillary electrophoresis, amino acid composition, se- Tyr-Lys . quence analysis, mass spectrometry, and NMR spectroscopy (13, 14, 23, The CPP stabilize calcium and phosphate ions under neutral 28 –31). and alkaline conditions forming metastable solutions that are Calcium Binding to  -CN(59–79), -CN(1–25),  -CN(63–70), and S1 S1 -CN(71–78)—All calcium binding experiments were performed using supersaturated with respect to the basic calcium phosphate S1 a modification of the ultrafiltration method of Marsh (32) without the phases (15). Under these conditions, the CPP bind their equiv- addition of phosphate. The solutions were buffered using 100 mM Tris- alent weights of calcium and phosphate (16). The CPP are HCl to a pH value of 7.0, 7.5, 8.0, 8.5, or 9.0. Ionic strength was adjusted formed in vivo by normal digestion of casein and, because they to 0.16 – 0.19 using NaCl. The reaction mixtures were incubated at room are relatively resistant to further proteolytic degradation, ac- temperature for 18 h. Subsequently, less than 10% of the total volume cumulate in the distal portion of the small intestine (17–21). It was collected as ultrafiltrate by centrifugation at 1000  g for 15 min using Centrifree MPS-1 Micropartition cells (Amicon) equipped with has been proposed that this accumulation together with the YM-3 (3000 molecular weight exclusion limit) or YM-1 (1000 molecular ability of the peptides to form soluble complexes with calcium weight exclusion limit) membranes. These membranes were demon- phosphate are responsible for the enhanced intestinal calcium strated not to retain calcium or phosphate ions or ion pairs. Calcium ion absorption that has been observed even in vitamin D-deficient concentrations of acidified samples were measured at 422.7 nm by animals consuming dietary CPP (17–21). In addition, CPP in- atomic absorption spectroscopy using PerkinElmer instrument model crease the calcification of in vitro cultured embryonic rat bone, 373, with the addition of 1% LaCl to prevent phosphate interference. Calcium concentrations ranged from 0.5 to 21.5 mM for  -CN(59 –79) and again the mechanism is suggested to be associated with the S1 and -CN(1–25), 0.3–5.5 mM Ca for the  -CN(63–70) peptide, and S1 ability of the peptide to form soluble complexes with calcium 0.05–1.5 mM Ca for the  -CN(71–78) peptide. The calcium concen- S1 and phosphate ions (22). Furthermore, CPP-calcium phosphate trations of the original reaction mixture after centrifugation to demon- complexes have been shown to be anticariogenic and to remi- strate no precipitation and of the ultrafiltrate were determined, and the neralize early stages of enamel caries in animal and human peptide-bound calcium was calculated as the difference between these studies (12, 15, 23–25). In summary, the ability to stabilize values. Calcium binding by the peptides was modeled by assuming the number of independent ion-binding sites/peptide ( ), where each site calcium phosphate and thereby enhance mineral solubility and Ca has a dissociation constant (K ) given by the following equation. bioavailability (26) confers upon the CPP the potential to be biological delivery vehicles for calcium and phosphate (27). Peptide][Ca ] As part of our long term investigation into the structure- K  (Eq. 1) [PeptideCa ] function relationships of proteins involved in biomineralization and calcium phosphate stabilization, we have studied the in- The dissociation constants for the peptide/calcium complexes (K ) and the number of calcium ion-binding sites/mol of peptide ( ) were de- teraction of tryptic phosphopeptides from milk caseins with the Ca termined by a nonlinear least squares fit to the equation, amorphous and crystalline phases of calcium phosphate. In this paper, we report our investigations of the calcium and [Ca ] Ca free phosphate binding properties of the two major CPP, -CN(1– [Ca ]  (Eq. 2) bound K  Ca ] free 25) and  -CN(59 –79). We demonstrate that the experimen- S1 tally determined ion activity product of a basic amorphous where K  K . Calcium Phosphate Binding to  -CN(59–79) and -CN(1–25) Ho- calcium phosphate (ACP) phase best correlates with the cal- S1 mologues and Analogues—Binding was initiated by adding phosphate cium bound by the peptide  -CN(59 –79) over a range of S1 to solutions containing 14 mM calcium and 4 mg/ml peptide buffered calcium and phosphate concentrations and sample pH ranging using 100 mM Tris-HCl to a pH value of 7.0, 7.5, 8.0, 8.5, or 9.0. Sodium from 7.0 to 9.0. Furthermore, we delineate the regions and chloride was added to bring the ionic strength for each sample to 0.16. residues of these peptides that are responsible for calcium For the homologues and analogues, binding was determined at pH 7.0 phosphate stabilization. We report the effect of peptide length, or 9.0. The reaction mixtures were incubated and ultrafiltered as de- scribed above for calcium binding. The phosphate concentration ranged residue type, and order of acidic residues on the binding to from0to9mM. Phosphate concentration was determined colorimetri- calcium and calcium phosphate using a library of synthetic cally (33), with absorbance measured at 660 nm on a PerkinElmer 552 analogues. Finally, we describe the ultrastructure of the casein Spectrophotometer. Calcium and phosphate concentrations in the orig- phosphopeptide-calcium phosphate nanocomplex as deter- inal solution and the ultrafiltrate were determined. The peptide bound mined using a range of physicochemical techniques including calcium and peptide bound phosphate were taken as the difference powder diffraction x-ray crystallography, scanning electron mi- between the total and free calcium and phosphate, respectively. To confirm that no precipitation had occurred during incubation, the sam- croscopy (SEM), and transmission electron microscopy (TEM). ples were centrifuged at 17,000  g for 5 min, and the calcium and EXPERIMENTAL PROCEDURES phosphate concentrations were determined prior to ultrafiltration. Ul- Preparation of Casein Phosphopeptides—The casein phosphopep- trafiltrates were examined spectroscopically at 214 nm and were found tides -CN(1–25) and  -CN(59 –79) were selectively precipitated from to contain no peptide. Peptide-bound calcium phosphate is expressed as S1 a tryptic digest of casein using calcium chloride and ethanol and further the number of calcium ions ([Ca ] ) and phosphate ions ([P ] )/ bound i bound purified by anion exchange fast protein liquid chromatography and mol of peptide. The ion activity products for various phases of calcium reversed phase HPLC (13). The purity of the peptides was assessed by phosphate were determined from the free calcium and phosphate con- matrix-assisted laser desorption ionization time-of-flight mass spec- centrations and pH using an iterative computational procedure that trometry, capillary electrophoresis, amino acid composition, and se- calculates the ion activity coefficients using the expanded Debye- quence analyses (13, 14). Prior to sequence analysis, the labile phos- Hu ¨ ckel equation. This procedure takes into account ion pairs CaHPO , phoseryl residues were converted to S-ethyl cysteinyl residues by CaH PO , and CaPO ; the dissociation of H PO and H O; and the 2 4 4 3 4 2 -elimination (13). ionic strength (32, 34). The activity of the CaOH ion was explicitly Preparation of Synthetic Peptides—The peptide Ac-Glu-Ser(P)-Ile- assumed to be negligible. Dissociation constants at 37 °C were used Ser(P) -Glu -NHMe corresponding to  -CN(63–70) was prepared by from the following sources: H PO (35); HPO (36); CaH PO and 3 2 S1 2 4 4 2 4 15364 Casein Phosphopeptide-Amorphous Calcium Phosphate CaHPO (37); and CaPO (38). The sources of the solubility products at 4 4 37 °C were hydroxyapatite (HA) (39), octacalcium phosphate (OCP) (40), dicalcium phosphate dihydrate (DCPD) (37), and ACP (41). Cal- cium and phosphate binding was modeled as occurring at multiple independent sites within the peptide with each site having a K de- d2 fined by the following equation. Peptide][Ca ][P ] K  (Eq. 3) d2 [PeptideCa P ] The number of phosphate binding sites was determined by nonlinear least squares fits to an equation of the form of Equation 2, in which bound and free phosphate were the variables. The value of K from these fits can be shown to be approximately given by the equation, d2 K  (Eq. 4) where K is the calcium ion binding constant for the peptide. The number of calcium binding sites was similarly determined using the equation, Ca i free 2 2 Ca    Ca  (Eq. 5) bound initial K  P i free FIG.1. Calcium ion binding by casein phosphopeptides. Titra- 2 tion of peptides with calcium ions. Bound calcium is plotted as a func- where [Ca ] is the amount of calcium bound with no added Initial tion of the free calcium ion concentration. Units of bound calcium are phosphate. calcium ions bound/mol of peptide. (f)  -CN(59 –79), () -CN(1–25), S1 Preparation of CPP-Calcium Phosphate—CPP was dissolved at 10 and (Œ) Ac-Glu-Ser(P)-Ile-Ser(P)-Ser(P)-Ser(P)-Glu-Glu-NHMe, the g/liter in Milli-Q water. 1.6 M CaCl and 1 M Na HPO were added 2 2 4 synthetic octapeptide corresponding to  -CN(63–70). The solid curves S1 slowly by syringe pump (0.5–1.0 ml/min) to the CPP solution in a are nonlinear least squares fits of the data to Equation 2. pH-stat held at pH 9.0 by automatic titration of 5 N NaOH. After 60 min of titration the final calcium and inorganic phosphate concentration TABLE I were 100 and 60 mM, respectively. The colloidal CPP-calcium phosphate Calcium binding characteristics of the casein phosphopeptides nanocomplexes were then concentrated 5-fold by microfiltration Peptide pH V K Ca d through a 0.1-m Sartorius filter (polyolefin) in a mini-Sarticon micro- b b filtration system and washed with 3 volumes of Milli-Q water to remove -Casein(63–70) 8.00 3.52  0.13 2.71  0.18 S1 free calcium and phosphate ions. The final CPP-calcium phosphate -Casein(59–79) 8.00 6.84  0.45 5.07  0.61 S1 solutions were then lyophilized. -Casein(1–25) 8.00 4.64  0.17 2.28  0.25 Preparation of -CN(1–25)-Calcium Phosphate—The -CN(1–25) v  number of calcium-binding sites/molecule of peptide. Ca phosphopeptide was dissolved at 1.67 g/liter in Milli-Q water. To the Means  S.D. (mM). peptide solution, 50 mM CaCl and 21.3 mM Na HPO were added 2 2 4 stepwise slowly. The pH was held at 9.0 by automatic titration of 50 mM Calcium Phosphate Binding to  -CN(59–79) and -CN(1– NaOH. The final -CN(1–25)-ACP solution was then lyophilized and S1 25)—Fig. 2 shows the titration of  -CN(59 –79) with inor- then redissolved in H O. S1 Powder Diffraction of CPP-ACP—The powdered samples were ganic phosphate in the presence of 14 mM calcium and at mounted in an aluminum holder and were placed in a Sintag pad V various pH values. Peptide-bound calcium and phosphate were x-ray diffractometer, operating with a Cu K (  1.5406 Å) x-ray determined by the difference in the total and free calcium and source at 45 kV and 40 mA. The x-ray diffraction data were collected phosphate concentrations, respectively. The quantity with a germanium solid state collector and scanned from 10° to 100° in 2 2 0 2 0 [Ca ]  [Ca ] , where [Ca ] is the calcium 2 at 2° min (step size 0.03°). Phase identification was accomplished bound bound bound bound in the absence of phosphate, represents the additional by comparison of the x-ray diffraction data pattern with the JCPDS powder diffraction file, produced by the International Center for Dif- calcium bound in the presence of phosphate. This quantity has fraction Data. been plotted versus free phosphate (P ) in Fig. 2A. Fig. 2B SEM and TEM of CPP-ACP—The sample for TEM was prepared shows a plot of bound phosphate ([P ] ) versus free P . i bound i from powdered material that was crushed and placed on a 3-mm TEM Nonlinear least squares fits were performed to determine the grid. The TEM grids were covered with a Butvar/Formvar (ProSciTec, number of mol of calcium and phosphate bound and the effec- Thuringowa Central, Australia) film to support the powder. Transmis- tive dissociation constant K. The fits to Equation 5 showed a sion electron micrographs were taken on a 200-kV JEOL 200CX instru- 2 0 ment operating at 200 kV. strong correlation between the value of [Ca ] and the bound 2 0 For SEM, specimens were sprinkled on an aluminum stub covered value of K.If[Ca ] was not constrained, it adopted un- bound with adhesive conductive carbon tabs and gold sputter-coated (Gold reasonable values, and the value of K determined from fits to Sputter coater S150B, Edwards, UK). Observations were made with a the phosphate data disagreed with those determined from the field emission SEM instrument (Philips XL 30 FEG, Eindhoven, The 2 0 calcium data. Fixing [Ca ] at the experimental value of Netherlands) operating at 20 kV using the secondary electron mode. bound the bound calcium ion concentration at zero phosphate gave RESULTS values of K in reasonable agreement with those derived from Calcium Binding to  -CN(59–79), -CN(1–25),  -CN(63– the phosphate data. The number of calcium ion-binding sites/ S1 S1 70), and  -CN(71–78)—Fig. 1 shows the calcium-binding mol of peptide ( ) was not sensitive to the value of S1 Ca 2 0 curves for (a)  -CN(59 –79) and (b)  -CN(63–70) with the [Ca ] . The values reported in Table II are appropriately S1 S1 bound bound calcium/mol of peptide ([Ca ] ) being plotted as a weighted averages of the experimental parameters determined bound function of free calcium ion concentration. Table I summarizes from repeated measurements. Fig. 2C, a plot of [Ca ] bound the values of the dissociation constants and the number of versus [P ] for the samples at different pH, reveals a linear i bound calcium ion-binding sites/peptide from fitting the data for  - relationship between bound calcium and bound phosphate. S1 CN(59 –79), -CN(1–25), and  -CN(63–70) to Equation 2 us- This demonstrates that the binding of calcium ions and phos- S1 ing a nonlinear least squares algorithm. In the case of the phate ions is co-dependent. The number of calcium phosphate C-terminal region  -CN(71–78) homologue containing only a binding sites/mol of peptide shows a strong correlation with the S1 single phosphoserine, no calcium binding was observed. length of the peptide as shown in Fig. 5. Furthermore, the ratio Casein Phosphopeptide-Amorphous Calcium Phosphate 15365 FIG.2. Calcium phosphate binding by  -CN(59–79). Titration of  -CN(59 –79) with inorganic phosphate in the presence of 14 mM S1 S1 calcium at various pH values is shown. A, bound calcium versus free phosphate. B, bound inorganic phosphate versus free P . C, bound calcium versus bound inorganic phosphate. of [Ca ] versus [Pi] increases as the peptides become z  0.00 to z  0.20. At each value of z, the ion activity product bound bound shorter (Table II). was plotted as a function of the number of calcium-binding Determination of the Calcium Phosphate Phase Stabilized by sites, and a quartic polynomial was fitted to the data. The -CN(59–79)—To determine the calcium phosphate phase correlation coefficients from these fits were then plotted S1 bound by  -CN(59 –79), the binding of calcium phosphate against z, and a quadratic polynomial fitted through these data S1 was related to the ion activity products of the various biolog- points to determine the optimal value of z. Fig. 4 shows the ically relevant calcium phosphate phases. The bound calci- variation in the correlation coefficient as a function of z for the um/mol of peptide at various pH values was plotted as a amorphous phases. function of the ion activity products of the following biologi- Calcium Phosphate Binding to Homologues and Ana- 2 2 cally relevant phases: DCPD (Ca )(HPO ) (Fig. 3A) (42), logues—In addition to -CN(1–25) and  -CN(59 –79), the cal- 4 S1 2  3 2 3 OCP (Ca ) (H )(PO ) (Fig. 3B), HA (Ca ) (PO ) (OH ) cium phosphate binding characteristics of a series of synthetic 4 4 3 5 4 3 2 3 (Fig. 3C), and a basic ACP (Ca ) (PO ) (OH ) , where homologues and analogues were investigated as shown in Ta- 3z 4 2 2z z  0.0877 (Fig. 3D). Only the ACP phase produced a function ble II. The aim of this investigation was to delineate the resi- that was independent of pH (Fig. 3D), indicating that this was dues important for calcium phosphate stabilization and to in- the phase stabilized by  -CN(59 –79). Two types of ACP were vestigate the importance of phosphorylation, peptide length, S1 considered, acidic ACP phases represented by (Ca ) - residue type, and order. The synthetic peptides, Ile-Val-Pro- 3z/2 3 2 (PO ) (HPO ) and basic ACP phases represented by Asn-Ser(P)-Glu-Glu and Gln-Met-Glu-Ala-Glu, corresponding 4 2z 4 z 2 3 (Ca ) (PO ) (OH ) . These formulae are capable of rep- to the C and N termini of  -CN(59 –79), respectively, did not 3z 4 2 2z S1 resenting a wide range of calcium phosphate compositions bind calcium or stabilize calcium phosphate. The importance of ranging from HA (z  ⁄3) in the basic range, to tricalcium the multiphosphorylated motif Glu-Ser(P)-Ile/Leu-Ser(P) -Glu 3 2 phosphate (z  0), OCP (z  ⁄3) and DCPD (z  2) in the acidic was further investigated using a series of analogues. A plot of range. To determine the optimum stoichiometry, ion activity  versus the number of residues in the peptide is shown in Ca products were calculated at a number of points in the interval Fig. 5, which produced a linear correlation coefficient of 0.9956. 15366 Casein Phosphopeptide-Amorphous Calcium Phosphate TABLE II Calcium and phosphate binding characteristics of CPP and analogues a b c Peptide pH Charge   K Calcium:P Ca Pi d i -Casein(59–79) 7.0 13.52 16.86  0.35 11.22  0.56 1.63  0.17 1.50  0.11 S1 7.5 13.85 23.2  1.6 15.8  1.1 0.74  0.10 1.46  0.20 8.0 13.99 21.73  0.68 14.26  0.38 0.257  0.021 1.524  0.088 8.5 14.10 22.3  2.3 14.3  1.6 0.064  0.020 1.56  0.33 9.0 14.29 21.43  0.39 14.05  0.79 0.072  0.016 1.52  0.11 -Casein(59–79) N74D 7.0 14.20 15.3  1.5 10.32  0.56 1.48  0.22 S1 9.0 15.27 22.35  0.33 14.25  0.57 1.569  0.086 -Casein(63–70) 9.0 11.00 9.64  0.26 5.60  0.14 0.132  0.015 1.721  0.089 S1 E I EE 9.0 10.99 7.85  0.22 5.11  0.15 1.538  0.090 -Casein(59–63) 9.0 2.00 Precipitated S1 -Casein(71–78) 9.0 2.99 Precipitated S1 -Casein(1–25) 7.0 12.61 15.65  0.16 10.55  0.98 1.48  0.15 9.0 13.26 24.76  0.63 16.67  0.47 1.485  0.079 -Casein(14–21) 9.0 11.00 12.01  0.30 6.35  0.16 0.084  0.024 1.891  0.095 Ser(P) -Glu 9.0 8.00 8.76  0.19 4.804  0.083 1.824  0.071 3 2 Thr(P) -Glu 9.0 7.99 4.90  0.10 2.69  0.16 1.82  0.16 3 2 Ser-Thr(P) -Glu 9.0 6.00 1.861  0.051 1.104  0.061 1.69  0.14 2 2 Ser-Ser(P) -Glu 9.0 6.00 1.711  0.084 1.16  0.12 1.47  0.22 2 2 Glu-Ser(P) -Glu 9.0 7.00 2.045  0.029 1.10  0.11 1.86  0.22 2 2 Asp-Ser(P) -Glu 9.0 7.00 1.917  0.088 0.992  0.052 1.93  0.19 2 2 (Ala-Ser(P) -Ala-Glu 9.0 9.00 4.948  0.049 2.899  0.065 1.670  0.037 (Ala-Ser(P)) -Ala-Glu 9.0 7.00 Precipitated (Ala-Ser(P)) -Ala-Glu 9.0 5.00 Precipitated Ala-Glu-Ala 9.0 4.00 Precipitated Ala-Glu-Ala 9.0 3.00 Precipitated Ala-Glu 9.0 3.00 Precipitated Charges calculated using the pK parameters listed in Bundi and Withrich (46) and Bienkiewicz and Lumb (47). b 2 number Ca -binding sites/mol of peptide. Ca number of P -binding sites/mol of peptide. Pi i Mean  S.D. (mM). E  Glu, I  Ile,  Thr(P). FIG.3. pH dependence of ion activ- ity products of calcium phosphate phases and peptide-bound calcium. The number of mol of bound calcium at various pH values plotted as a function of the activity product of the DCPD 2 2 phase (Ca )(HPO )(A), the OCP phase 2 3 (Ca ) (PO ) (H )(B), the HA phase 4 4 3 2 3 (Ca ) (PO ) (OH )(C), and the basic 5 4 3 2 3 ACP phase (Ca ) (PO ) (OH ) 3.0877 4 2 0.1754 (D). The curves in A–C are interpolated curves using the method of Stineman (42) provided as an aid to visualization and to emphasize that a single function does not relate the bound calcium to the particular ion activity product. The curve in D is a quartic polynomial fit to all the data points. Casein Phosphopeptide-Amorphous Calcium Phosphate 15367 FIG.6. Powder diffraction spectra of CPP-ACP complexes. Shown are the X-ray powder diffraction spectra of CPP-ACP (A) and heat-treated CPP-ACP (B) showing peaks characteristic of an apatitic phase. The vertical bars show the expected positions of peaks in the x-ray powder spectrum of hydroxyapatite. phosphate bound by the CPP and their analogues increased as the pH increased. The calcium phosphate phase that was found to bind to  -CN(59 –79) was a basic ACP phase of the com- S1 position, Ca (PO ) (OH) . 3.0877 4 2 0.1754 X-ray Powder Diffraction of CPP-ACP—A number of samples FIG.4. Correlation of bound calcium and ion activity products of CPP-ACP and -CN(1–25)-ACP, were studied using x-ray pow- as a function of calcium phosphate phase. Plot showing the vari- ation in correlation coefficient as a function of phase composition (z) for der diffraction. An understanding of the degree of the crystallin- quartic polynomial fits to the number of calcium ions bound/mol of ity of the CPP-ACP complex can be obtained from the line broad- peptide as a function of the ion activity products of the acidic (dashed ening of the peak intensity spectrum for the polycrystalline line) and basic ACP (solid line) phases. powder. A sample spectrum is shown in Fig. 6A. The absence of sharp features in these spectra was consistent with an amor- phous calcium phosphate phase associated with the CPP and the individual peptides. A few peaks in the spectrum of the -CN(1– 25)-ACP complex were observed and assigned to traces of crys- talline NaCl and KCl contaminating the sample. Heat treatment of CPP-ACP produces material with less bio-available calcium and phosphate, and x-ray powder diffraction patterns indicated some partial crystallinity with conversion to a disordered apatitic phase upon heat treatment, as shown in Fig. 6B. SEM and TEM Characterization of CPP-ACP—Neither SEM nor TEM showed any crystallites in the CPP-ACP samples. Selected area electron diffraction over the particles also did not show any ring pattern characteristic of crystal structure. The SEM and TEM images of CPP-ACP and the selected area diffraction pattern therefore confirm the amorphous nature of the complex (Fig. 7). DISCUSSION FIG.5. Correlation of bound calcium to peptide length in cal- cium phosphate-peptide complexes. Plot showing correlation be- Calcium Binding—As shown in Table I,  -CN(59 –79) max- S1 tween the number of residues in the peptide and the number of calcium imally bound 6.84  0.45 mol of Ca /mol of peptide, whereas ion-binding sites/mol of peptide for the CPP-ACP complexes and their 2 -CN(1–25) bound 4.64  0.17 mol of Ca /mol of peptide. The analogues. The linear fit has a correlation coefficient of 0.993. shorter peptide  -CN(63–70) bound 3.52  0.13 mol of Ca / S1 mol of peptide. All of the measurements were performed at pH The substitution of any of the Ser(P) residues by Ser, Glu, or 8. The number of calcium ion-binding sites ( ) appears to be Ca Asp residues in the peptide Ser(P) -Glu or the substitution of 3 2 determined by charge for the longer peptides, the amount of Thr(P) residues by Thr in the Thr(P) -Glu peptide substantially calcium bound being sufficient to nearly cancel the intrinsic reduced the peptides ability to stabilize calcium phosphate. negative charge of the peptides. Further, the substitution of Ser(P) by Thr(P) was also associ- The charge on the peptide  -CN(63–70) is only slightly S1 ated with a reduced ability to stabilize calcium phosphate. lower than that on the other two peptides, and yet the amount Three contiguous phosphoseryl residues were necessary for full of calcium bound is significantly lower than would be expected activity, because the (Ala-Ser(P)) peptide, although containing based on charge neutralization. A structural explanation ap- four noncontiguous phosphoseryl residues, poorly stabilized pears to be the most plausible for the markedly lower than calcium phosphate. expected calcium binding ability of the shorter peptide. Inspec- In summary, in the presence of phosphate ions, the CPP and tion of calcium-binding proteins whose structures have been their analogues bound additional calcium (over and above that deposited in the Protein Data Bank suggests that calcium ions bound in the absence of phosphate). The linear dependence of the excess calcium bound versus the bound phosphate, inde- pendent of pH, implies that a specific phase of calcium phos- K. J. Cross, N. Laila Huq, J. E. Palamara, and E. C. Reynolds, phate was being bound (Fig. 2C). The amount of calcium and unpublished observations. 15368 Casein Phosphopeptide-Amorphous Calcium Phosphate calcium phosphate phase. Once the peptide is bound to the ACP phase, weaker interactions allow the remaining polar residues to interact with the ACP core particle, thus accounting for the correlation between peptide length and the number of mol of calcium and phosphate bound by these peptides. Given that the entire length of the peptide interacts with the calcium phos- phate phase and the stringent steric requirements associated with the formation of calcium ion-binding sites, it is unlikely that peptides separately bind calcium ions that are not associ- ated with the calcium phosphate phase. Characterization of the Bound Calcium Phosphate Phase— Fig. 4 shows plots of the calcium bound by the peptide -CN(1– 25) versus ion activity products for the four relevant calcium phosphate phases: HA, DCPD, OCP, and a basic amorphous calcium phosphate having the composition Ca (PO ) (OH) 3z 4 2 2z where z  0.0877. The calcium phosphate phase stabilized by -CN(1–25) is identified by the functional dependence of bound calcium (or phosphate) on the ion activity product independent of sample pH. Inspection of Fig. 4 shows that the amount of calcium bound is not a pH-independent function of the ion activity products for the phases HA, DCPD, or OCP. Fig. 4 shows that the bound calcium best correlates with a basic amorphous calcium phosphate phase Ca (PO ) (OH) , with 3z 4 2 2z z  0.0877  0.0022. The stabilization of a basic calcium phosphate phase is consistent with the observation that these complexes form over a pH range from 5.0 to 9.0. The ratio of  : was found to increase as the length of the Ca Pi peptide decreased (Table II). One possible explanation for this FIG.7. Electron microscopy of CPP-ACP complexes. a, SEM observation would be that the shorter peptides bind both a image of CPP-ACP; the segments of the scale bar represent a distance of 10 m. b, TEM image of CPP-ACP at 100,000. c, selected area calcium phosphate phase and independent calcium ions. How- diffraction pattern of CPP-ACP. Note the absence of evidence for crys- ever, as discussed above, the stringent requirements associated talline material. with the formation of calcium ion-binding sites makes this an unlikely explanation. usually have between six and eight oxygen atoms in their first These separate observations can be rationalized by propos- coordination sphere. The structure of the calcium-binding sites ing a model of the ACP core that consists of two calcium is based on distorted octahedral geometries with “small bite” phosphate phases: a calcium-poor phase with a Ca:P ratio of ligands, such as carboxylate groups, contributing two oxygen 1.5, such as Ca (PO ) , forming the core of the ACP particles as 3 4 2 atoms at some vertices of the octahedra. Water molecules usu- suggested by Meyer and Eanes (45), and a calcium-rich phase ally account for only one or two oxygen atoms/binding site. with a Ca:P ratio of 2.0, such as Ca (PO )(OH), that is in 2 4 Apart from interactions with side chain oxygen atoms, calcium contact with the peptide. As the peptides become shorter, a ions also interact with oxygen atoms from the backbone car- given number of peptides are able to fully cover a smaller bonyl moieties. The structural constraints implied by the mul- surface area; hence the size of the complex formed decreases as tiple interactions between the shorter peptide and the calcium the peptides become shorter. The thickness of the ions that are required to balance the peptide charge are likely Ca (PO )(OH) phase interacting with peptide is proposed to 2 4 to be impossible to satisfy, resulting in the lower calcium bind- remain constant as the complexes shrink in size, whereas the ing ability of the shorter peptide. The structural explanation is size of the relatively more soluble Ca (PO ) core decreases in 3 4 2 consistent with our previous reports of calcium ion-dependent, size, resulting in a steady increase in the ratio of  : as the Ca Pi structural features in the H NMR spectra of  -CN(59 –79) S1 peptides decrease in length. It is interesting that substitution and -CN(1–25) complexed with calcium (43, 44). We have of Ser(P) with Thr(P) reduced calcium phosphate stabilization shown that these peptides adopt structures consisting of loops activity. This together with possible steric hindrance in phos- and turns and that the specific structure adopted depends on phorylation of Thr residues may help explain why to date all of the peptide sequence outside of the calcium-binding motif the proteins known to stabilize calcium phosphate contain clus- Ser(P) -Glu . 3 2 ters of Ser(P) residues. Calcium Phosphate Binding—In the presence of phosphate The ion product that best correlates with the total bound ions, the CPP and their analogues bound additional calcium calcium by  -CN(59 –79) in a pH-independent manner is that (over and above that bound in the absence of phosphate). The S1 of a basic amorphous calcium phosphate phase having the linear dependence of the excess calcium bound versus the approximate composition Ca (PO ) (OH) . Given that bound phosphate, independent of pH, implies that a specific 3.0877 4 2 0.1754 each  -CN(59 –79) maximally binds 14 phosphate ions, the phase of calcium phosphate was being bound (Fig. 2C). The S1 composition of the  -CN(59 –79)-ACP complex is approxi- amount of calcium and phosphate bound by the CPP and their S1 mately ( -CN(59 –79)-(ACP) ) , where n is a small integer. analogues increased as the pH increased. S1 7 n Extensive studies using a library of synthetic homologues The number of calcium-binding sites in the calcium phos- phate complexes ( ) correlates strongly with the number of and analogues have revealed that the Ser(P) -Glu motif found Ca 3 2 in most of the CPP is a specific sequence of acidic residues that aminoacyl residues in the peptide used to stabilize the CPP- ACP complex. This is consistent with a model of the CPP-ACP has a strong affinity for calcium phosphate. However, the con- formational preferences of the intact major peptides enable complexes in which the phosphorylated sequence motif, Ser(P)- Ser(P)-Ser(P)-Glu-Glu, is required to initialize binding to the stabilization of the maximum amount of ACP. Casein Phosphopeptide-Amorphous Calcium Phosphate 15369 206 –211 Acknowledgment—We thank Fiona Webber for technical assistance. 25. Shen, P., Cai, F., Nowicki, A., Vincent, J., and Reynolds, E. C. (2001) J. Dent. REFERENCES Res. 80, 2066 –2070 26. Kitts, D. D., Yuan, Y. V., Nagasawa, T., and Moriyama, Y. (1992) Brit. J. Nutr. 1. Walstra, P., and Jenness, R. (1984) Dairy Chemistry and Physics, John Wiley 68, 765–781 & Sons, Inc., New York 27. Ferraretto, A., Signorile, A., Gravaghi, C., Fiorilli, A., and Tettamanti, G. 2. Van Hooydonk, A. C. M., Boerrigter, I. J., and Hagedoorn, H. G. (1986) Neth. (2001) J. Nutr. 131, 1655–1661 Milk Dairy J. 40, 297–313 28. Perich, J. W., Johns, R. B., and Reynolds, E. C. (1992) Aust. J. Chem. 45, 3. Holt, C., and Sawyer, L. (1988) Protein Eng. 2, 251–259 385–394 4. Schmidt, D. G. (1982) Dev. Dairy Chem. 1, 61– 86 29. Perich, J. W., Kelly, D. P., and Reynolds, E. C. (1992) Int. J. Pept. Prot. Res. 40, 5. McGann, T. C., Kearney, R. D., Buchheim, W., Posner, A. S., Betts, F., and 81– 88 Blumenthal, N. C. (1983) Calcif. Tissue Int. 35, 821– 833 30. Perich, J. W. (1991) Method Enzymol. 201, 225–233 6. McGann, T. C., Buchheim, W., Kearney, R. D., and Richardson, T. (1983) 31. Perich, J. W., Terzi, E., Carnazzi, E., Seyer, R., and Trifilieff, E. (1994) Int. J. Biochim. Biophys. Acta 760, 415– 420 Pept. Prot. Res. 44, 305–312 7. de Kruif, C. G. (1999) Int. Dairy J. 9, 183–188 32. Marsh, M. E. (1989) Biochemistry 28, 346 –352 8. Holt, C., and Horne, D. S. (1996) Neth. Milk Dairy J. 50, 85–111 33. Itaya, K., and Ui, M. (1966) Clin. Chim. Acta 14, 361–366 9. Horne, D. (1998) Int. Dairy J. 8, 171–177 34. Smales, F. C. (1972) Calcif. Tissue Res. 8, 304 –319 10. Dalgleish, D. G., Spagnuolo, P. A., and Goff, H. D. (2004) Int. Dairy J. 14, 35. Bates, R., and Acree, S. (1943) J. Res. Natl. Bureau Stand. 30, 129 1025–1031 36. Bjerrum, N. (1929) Kgl. Danske. Videnskab. Selskab. Math. fys. Medd. 9, 5–206 11. Holt, C., Wahlgren, N. M., and Drakenberg, T. (1996) Biochem. J. 314, 37. Gregory, T. M., Moreno, E. C., Brown, W. E. (1970) J. Res. Natl. Inst. Stand. 1035–1039 74A, 461 12. Reynolds, E. C., Black, C. L., Cai, F., Cross, K. J., Eakins, D., Huq, N. L., 38. Chughtai, A., Marshall, R., and Nancollas, G. H. (1968) J. Phys. Chem. 72, Morgan, M. V., Nowicki, A., Perich, J. W., Riley, P. F., Shen, P., Talbo, G., 208 –211 and Webber, F. (1999) J. Clin. Dent. 10, 86–88 39. McDowell, H., Gregory, T. M., and Brown, W. E. (1977) J. Res. Natl. Bur. 13. Reynolds, E. C., Riley, P. F., and Adamson, N. J. (1994) Anal. Biochem. 217, Stand. Sect. A. 81A, 273–281 277–284 40. Shyu, L. J. (1982) in The solid/solution interface-a kinetic study of the crystal- 14. Adamson, N. J., Riley, P. F., and Reynolds, E. C. (1993) J. Chromatogr. 646, lization of calcium fluoride and phosphate. Ph.D. thesis, State University of 391–396 New York, Buffalo, NY 15. Reynolds, E. C., Cain, C. J., Webber, F. L., Black, C. L., Riley, P. F., Johnson, I. H., and Perich, J. W. (1995) J. Dent. Res. 74, 1272–1279 41. Nancollas, G. H., LoRe, M., Perez, L., Richardson, C., and Zawacki, S. J. (1989) Anat. Rec. 224, 234 –241 16. Reeves, R. E. (1958) Science 128, 472 17. Mykkanen, H. M., and Wasserman, R. H. (1980) J. Nutr. 110, 2141–2148 42. Stineman, R. W. A. (1980) Creative Computing 6, 54 –57 43. Huq, N. L., Cross, K. J., and Reynolds, E. C. (1995) Biochim. Biophys. Acta 18. Lee, Y. S., Noguchi, T., and Naito, H. (1983) Brit. J. Nutr. 49, 67–76 19. Lee, S. L., and Veis, A. (1980) J. Pept. Prot. Res. 16, 231–232 1247, 201–208 20. Meisel, H., and Fristar, H. (1988) Biol. Chem. Hoppe-Seyler. 369, 1275–1279 44. Cross, K. J., Huq, N. L., Bicknell, W., and Reynolds, E. C. (2001) Biochem. J. 21. Sato, R., Noguchi, T., and Naito, H. (1986) J. Nutr. Sci. Vitaminol. 32, 67–76 356, 277–285 22. Gerber, H. W., and Jost, R. (1986) Calcif. Tissue Int. 38, 350 –357 45. Meyer, J. L., and Eanes, E. D. (1978) Calcif. Tissue Res. 25, 59–68 23. Reynolds, E. C. (May 14, 1991) U. S. Patent 5,015,628 46. Bundi, A., and Wu ¨ thrich, K. (1979) Biopolymers 18, 285–297 24. Reynolds, E. C., Cai, F., Shen, P., and Walker, G. D. (2003) J. Dent. Res. 82, 47. Bienkiewicz, E. A., and Lumb, K. J. (1999) J. Biomolec. NMR 15, 203–206

Journal

Journal of Biological ChemistryAmerican Society for Biochemistry and Molecular Biology

Published: Apr 15, 2005

There are no references for this article.