Calcif Tissue Int (1992) 51:143-150
Calcified Tissue International 9 1992 Springer-Verlag New York Inc.
Competitive Adsorption of Magnesium and Calcium Ions onto Synthetic and Biological Apatites T. Aoba, 1 E. C. Moreno, 1 and S. Shimoda 2 1Forsyth Dental Center, 140 Fenway, Boston, MA 02115; and 2Tsurumi University School of Dentistry, Yokohama 230, Japan Received September 16, 1991, and in revised form November 4, 1991
Summary. Magnesium (Mg) is a conspicuous constituent of hard tissues but its possible role in biomineralization is poorly understood. It is possible that Mg 2+ adsorbed onto bioapatites may contribute to the modulation of crystal growth as such inhibitory activity has been reported for synthetic apatites. The present study was undertaken to determine the adsorption isotherms of Mg ions onto synthetic apatites and biominerals in tooth and bone tissues in the presence of other ions of natural occurrence. Synthetic crystals used as adsorbents were hydroxyapatite and, as a better prototype for the biomineral, Mg-containing carbonatoapatite. Human enamel and dentin materials were obtained from extracted, caries-free, permanent teeth. Porcine dentin materials at two developmental stages were obtained from erupted deciduous and unerupted permanent teeth of a 6-month-old slaughtered piglet. Porcine bone was obtained from the cortical portion of the mandible of the same animal. All biomineral samples were pulverized and then treated by plasma ashing (deproteination) at about 60~ Each of the powdered samples was equilibrated in solutions containing various initial concentrations of Mg 2+, Ca 2+, and Na § (or K +) as nitrate salts. Following equilibration, concentrations (and activities) of magnesium and calcium ions in the experimental solution were determined. The pH values of the equilibrium solutions were in the range of 6.2-6.5. Experimental data of the Mg adsorption onto hydroxyapatite were interpreted on the basis of a Langmuir-type model for binary systems assuming competition of Mg 2§ and Ca 2 + for the same adsorption sites on the crystal surfaces of the apatites. According to this model, the adsorbed Mg is expressed as a function of the ionic activity ratio (Mg 2 +)/(Ca 2 § in the equilibrium solution. The model contains two parameters, the adsorption selectivity constant Ks and the maximum number of adsorption sites N (ixmol/g). The numerical values of Ks were similar for all adsorbents used (synthetic and biological) and indicated the preferential adsorption of Ca 2 § probably due to spacial restrictions extending to the very surface of the crystals. The initial level of Mg 2 § in the surface pool was different in the various biominerals, probably reflecting the composition of fluid in which the biominerals were formed. Whereas the surface pool of Mg of human enamel was marginal, only 5% of the total Mg, significant fractions of the total Mg in human and porcine dentins (about 2030%), and porcine bone (about 40%) existed on the crystal surfaces. There were significant differences in the total Mg and the value of the parameter N between young (unerupted) and mature (erupted) dentin minerals. It was ascertained that
Offprint requests to: T. Aoba
the occupancy of adsorption sites by Mg ions became greater with maturation of the dentin tissues. The overall results suggest that the Mg-mineral interaction in tooth and bone tissues may be a highly tissue-specific process, presumably reflecting differences in fluid composition (particularly Ca and Mg activities) responsible for biomineralization. Key words: Adsorption - Magnesium - Calcium - Apatite crystals - Enamel - Dentin - Bone.
Various reports [1-3] suggest that pyrophosphate (PPi), citrate, and magnesium (Mg) are potent inhibitors of calcium phosphate precipitation in biological fluids, such as plasma ultrafiltrate and urine. It was also established that, for equimolar concentrations, PPi is the most effective inhibitor, followed by citrate and magnesium in decreasing order of effectiveness [1]. However, when the inhibitors are used at physiological concentrations, Mg (e.g., about 0.8 mM in plasma ultrafiltrate) provided the most significant inhibition of seeded crystal growth of hydroxyapatite [1, 4, 5]. Most likely, the inhibitory activity is related to the adsorption of these ions onto growth sites on the surface of the apatite crystals. Indeed, we reported [6] that the inhibitory activity of PPi on the apatite crystal growth can be explained on the basis of its adsorption onto the seed crystals, which is described by the Hill-de Boer adsorption model [7]. However, to our best knowledge, no distinct model for the adsorption of Mg onto apatite crystals has been reported. As expected from the ubiquitous existence of Mg in various hard tissues, fluids separated from mineralizing zones have significant concentrations of Mg ions. Of particular interest are the differences in Mg concentration observed in different tissues. Thus, the cartilage fluid separated from chicken hypertrophic zone [8] and the enamel fluid from developing enamel tissue [9] have Mg concentrations of about 0.9 mM, similar to the Mg concentration that provides an effective inhibition of calcium phosphate precipitation [1, 4, 5], whereas the Mg concentration in predentin fluid [10] is much lower, about 0.2 mM. Although the differences observed in Mg concentration in the fluids of calcifying tissues suggest differences in the roles played by Mg ions in biomineralization from tissue to tissue, the limited knowledge on this subject precludes the assignment of specific function(s) for Mg in the mineralization of enamel and cartilage. Furthermore, in spite of the abundant information on the total Mg contents in tooth and bone tissues, our knowledge is still limited as to its status, whether it is present as a free ion or organic-bound in the fluid phase, adsorbed onto crystal surfaces, or incorporated into the crystal lattice of the
144 mineral. Q u a n t i t a t i v e i n f o r m a t i o n as to t h e labile (in solution), surface, a n d stable pools o f M g ions in t o o t h a n d b o n e tissues is e s s e n t i a l to (1) a s s e s s p o s s i b l e f u n c t i o n a l c o n t r i b u tions o f M g ions in t h e r e g u l a t i o n o f b i o m i n e r a l i z a t i o n a n d (2) d e t e r m i n e t h e n a t u r e a n d p r o p e r t i e s (e.g., crystalline lattice s t o i c h i o m e t r y a n d solubility) o f the f o r m e d b i o m i n e r a l . Furt h e r m o r e , it is p o s s i b l e t h a t t h e r e p o r t e d [3-5] c r y s t a l g r o w t h inhibition b y M g 2+ as well as its stabilization of m e t a s t a b l e calcium p h o s p h a t e s [11, 12] is r e l a t e d to a d s o r p t i o n of the ion o n t o specific sites o f t h e solid p h a s e . A similar role m a y b e p l a y e d b y the M g 2+ o n crystal s u r f a c e s in b i o m i n e r a l i z a tion p r o c e s s e s . T h e first step in t h e t e s t o f this h y p o t h e s i s , u n d e r t a k e n in this p u b l i c a t i o n , is to d e t e r m i n e t h e a d s o r p t i o n o f M g 2+ o n t o well-defined, s y n t h e t i c apatitic crystals, particularly in t h e p r e s e n c e o f C a 2 + ions as t h e l a t t e r are ubiquitous in m i n e r a l i z a t i o n z o n e s . W e also r e p o r t a m o d e l for the a d s o r p t i o n o f M g z+ o n t o d e n t a l e n a m e l , d e n t i n , a n d b o n e minerals. Finally, w e outline a p r o c e d u r e t h a t , using simple a d s o r p t i o n m o d e l s , p e r m i t s t h e d e t e r m i n a t i o n of the surface p o o l o f M g 2+ in t h e b i o a p a t i t e s .
Materials and Methods
Synthetic Apatites, Enamel, Dentin, and Bone Synthetic crystalline apatite samples used were hydroxyapatite, CaOHAp, having a composition of Ca 37.5%, total P 18.7%, Mg 0.00%, and CO 3 0.0%. A better prototype of biomineral was a magnesiumcontaining carbonatoapatite, (Ca,Mg)-CO3Ap, prepared in our laboratories, with a composition of Ca 35.2%, total P 17.7%, Mg 0.39%, and CO3 3.0% [13]. Specific surface areas of OH-Ap and (Ca,Mg)CO3AP, determined by N 2 adsorption (Quantachrome, NY), were 26.5 and 6.7 m2/g, respectively. Biomineral samples used were human enamel and dentin, and porcine dentin and bone. Sampling procedures of human enamel and dentin were reported previously in detail [14]. Briefly, human, caries-free permanent premolars and molars, extracted for orthodontic or periodontal reasons, were collected and stored in deionized water containing 0.5% (v/v) chloroform at 4~ Prior to preparation of the enamel samples, the enamel surface was thoroughly cleaned with pumice powder and a rotary rubber cup. The underlying dentin was ground off under running water; the grinding was advanced well beyond the dentin-enamel junction. For collection of dentin tissue, sound roots of the teeth were separated and then enough exterior surface (as well as interior surface facing the pulp cavity) was removed using a dental burr under water. The prepared hollow crown (enamel) and dentin specimens were placed separately in a sonicator with water (to remove fine debris) and dried at room temperature. Porcine materials were obtained from a 6-month-old slaughtered piglet. Dentin was dissected from erupted deciduous and unerupted permanent teeth; in this study, no special care was given to possible local differences in Mg content within enamel or dentin (apical or incisal aspects). Dentin tissue from each of the tooth groups was cut in the form of blocks with a dental discoid blade, cooling with running water. The dissected dentin blocks were treated in a similar way described for the human sample. For preparation of the porcine bone sample, the cortical portion of the mandible was cut with a circular saw, cooling with running water. The dissected bone block was transferred into liquid nitrogen and then crushed into small pieces while frozen. Each of the enamel, dentin, and bone samples was pulverized with an agate mortar and pestle. The resulting powder was passed through a 270 mesh sieve and stored at - 20~ until used. Prior to the adsorption experiments, each pooled sample was ashed in a plasma (oxygen activated by a radio frequency) at about 60~ as reported previously [15]. The weight loss of the sample was monitored and the treatment was continued until a constant weight was obtained, which usually occurred within 48 hours. Following the treatment, the ashed sample (about 100 mg) was placed in a centrifuge tube. In order to wash out debris of decomposed organic
T. Aoba et al.: Adsorption of Mg z+ onto Apatite Crystals matter, one ml of deionized water, precooled on ice, was added to the tube, and the resulting slurry was dispersed with a sonicator. The suspension was separated by centrifugation at 10,000 x g for 10 minutes. The supernatant was carefully removed, and the sediment was redispersed with the addition of another 1 ml of water. The same procedure was repeated three times. Following the last washing, the sediment was freeze-dried and then used as required in the adsorption experiments. Volumes of the recovered supernatants were determined by weighing (assuming unitary density), and aliquots of the supernatants were used for analysis of Mg ions displaced from the solid in the solution (see below).
Adsorption of Mg e + onto Synthetic Apatites and Biominerals ACS-grade reagents Mg(NO3)26H20, Ca(NO3)24HzO, NaNO3, and KNO3 (Aldrich Chem Co., Milwaukee, WI) were used to prepare stock solutions. In the systems in which the adsorbents were synthetic apatites, the initial concentrations of magnesium and calcium in the experimental solutions were varied in the range of 0.0-10 raM. A total of 19 systems were used with Ca-OHAp and 15 with MgCO3Ap. Most of the adsorption experiments were conducted in the presence of 160 mM of either NaNO3 or KNO3 to provide an ionic strength similar to that found in physiological fluids. Some of the experiments (seven systems) were conducted at various concentrations of Na + ions (0, 50, and 100 mM), in order to investigate the effect of monovalent cations on the Mg adsorption onto the crystals. Prior to each experimentation, experimental solutions (50-100 ml total volume) having specified concentrations of total Mg and Ca were prepared from the stock solutions and then pure nitrogen was bubbled through them to expel CO2. Accurately weighed amounts of the crystals (in most cases, 10 mg) were added to 1 ml of the experimental solution. The resulting slurry was mixed initially with the aid of an ultrasonicator and then with a vortex mixer, followed by endover-over rotation at 25(-+ 1)~ In preliminary studies, with 10-100 mg of synthetic crystals added to 1 ml of the experimental solutions, Mg concentrations reached plateau values within 1 hour and did not change further over a 24-hour period. Therefore, the solid-solution equilibration in the following experiments was conducted for 2 hours unless otherwise specified. After equilibration, the suspension was centrifuged at 10,000 z g for 10 minutes. The recovered supernatant, as well as an aliquot of the original experimental solution, was used for chemical analyses (see below). The sedimented solid was freeze-dried and then used for analyses by X-ray diffraction (XRD), Fourier-transform infrared spectroscopy (FTIR), and electron microscopy (EM). Procedures for the instrumental analyses were reported previously in detail [13, 15]. In studies with each of the biomineral samples, a weighed powder sample (10 rag) was equilibrated in 1 ml of the experimental solution containing 160 mM NaNO 3. Initial concentrations of the total Mg and Ca were adjusted in the ranges of 0-3 mM and 0-1 mM, respectively. The rest of the experimental procedures were the same as described above. Eight systems were used with each of the human enamel, human dentin, young porcine dentin, and bone minerals; seven systems were used with older porcine dentin as the adsorbents.
Chemical Analysis The total Mg and Ca concentrations were measured by atomic absorption (AA) spectrophotometry. Sodium and potassium were determined with the same instrument in the emission mode. Standard solutions for Ca determination were prepared with CaCO3 (dissolved in dilute HCI solution); standard solutions for AA analyses of Mg, Na, and K were obtained commercially (Fisher Sci, Pittsburgh, PA). The total phosphorus was determined colorimetrically [16]. The estimated errors of these chemical determinations were as follows: Mg and Ca 3%, total P 1.5%, and Na and K 4%. The pH value of the experimental solution was determined using a glass-reference combination electrode (Fisher) and a pH meter (Orion, model 901, Cambridge, MA). Carbonate concentrations of the supernatant sam-
T. Aoba et al.: Adsorption of Mg2+ onto Apatite Crystals
145
pies were determined by ion chromatography (Dionex, Sunnyvale, CA) according to the method of Kreling and DeZwaan [17]. Based on the analytical data, the Mgz+ and Caz+ activities in the equilibrium solution were calculated using a generalized program, previously reported [18], that takes into account the formation of Mg and Ca ion-pairs with various forms of phosphate and carbonate ions.
Mg in the solution. Thus, the quantity ~Mg may be positive or negative (desorption of Mg2+), depending on the experimental conditions. Thus, equation (5) is rewritten as follows: R/(Q~ + 8Mg) = R/N + 1/K~N Rearranging the above equation, one obtains Qi § ~Mg = NKsR/(1 + KsR)
Adsorption Model of Mg Ions onto Apatite Crystal Surfaces Preliminary experiments indicated that adsorption of Mg was affected by the presence of Ca ions in the equilibrating solution and suggested a one-to-one exchange reaction on the apatitic surface. Consequently, the adsorption model adopted here considers a competing adsorption of Mg 2+ and Ca 2+ ions onto the apatite crystals (i.e., the two ions compete for the same adsorption sites). The equilibrium between ions in solution and on the crystal surface can then be expressed by the reaction C a ~ f ~ + Mg2+solution~-Mgsurface + CaZ+solution
(1)
Assuming that the exchangeable pool (i.e., the adsorption sites on the surface) accommodates a total of N moles of Ca2+ and Mg2+ per unit of surface area (or unit mass in the case of the biominerals), the equilibrium constant K for reaction (1) is given by K = QMg * (CaZ+)/[(N - QMg) * (Mg2+)]
(2)
in which parentheses represent activities of the ions enclosed and QMg stands for the moles of Mg2+ adsorbed per unit surface area or per unit of mass of adsorbent (the corresponding quantity for Ca2+ is given by N - QM~). Rearranging the above expression and making R = (Mg2+)/(Ca~+), the condition for equilibrium can be written as R/QMg = R/N + I/KN
(3)
This final equation is of the same form as that describing a Langmuir adsorption model, although the ratio R of Mgz+ and Caz+ activities is used here, instead of the Mg2+ activity in the equilibrated solution. The same equation can be derived by the use of a Langmuir adsorption model for a binary system [19] Qr~g = NKMg(MgZ+)/[1 + KMg(Mg2+) + Kca(Ca2+)] Qca = NKca(Ca2+)/[1 + KMg(Mg2+) + Kca(Ca2+)] N = QMg + Qca
(4)
in which KMg and Kc~ are the adsorption affinity constants of Mg2+ and Caz+ ions, respectively, onto the crystal surfaces. Calling Ks = KMg/Kc~ (the selectivity constant), combination of the three equations in (4) yields R/QMg = R/N + 1/KsN
(5)
which is formalistically identical to equation (3). If the model described by either equation (3) or (5) is correct, a plot of the experimental data in terms of R/QMg v e r s u s R should yield a linear relationship, and the numerical values of Ks and N can be obtained from the slope and intercept. Experimentally, numerical values of R were obtained based on the solution composition at the end of equilibration, whereas the corresponding QMg values were obtained by the difference in the total concentration ofMg 2+ ions in the experimental solutions before and after equilibration. In systems using Mg-containing apatite or the biominerals as adsorbents, it is reasonable to assume that part of the total Mg in the solid sample exists initially on the crystal surfaces in the same fashion as part of the Ca in Ca-OHAp is on the surface and exhibits exchange properties. Consequently, the above adsorption model should be modified to accommodate the initial surface pool of Mg ions, Qi. Calling gMg the adsorbed Mg experimentally determined after a given equilibration, the quantity Q~g (i.e., the total amount of Mgz+ in the adsorbed state on the crystals at equilibrium) is given by the sum of Qi and gMg. There is no a priori knowledge of the quantity Qi; also, the equilibrating solution may become enriched or impoverished in Mg2+ , depending on the ionic activities of Ca and
(6)
(7)
In equation (7), the three parameters N, Ks, and Qi are unknown. Their numerical values were obtained by the use of a nonlinear least-squares procedure together with the experimental data obtained from all the systems having a common adsorbent. The values of the three parameters are expected to be different for different adsorbents. Of particular interest are the values of Qi, the initial Mg present in the surface pool, and the value of Ks which, to some extent, should reflect the discrimination of Mg from the structure of Ca-apatites. In this publication, the term "adsorption isotherm" refers to the plot of the total amount of Mg in the surface pool Q~g (Qi plus the experimental quantity ~Mg) versus R (i.e., versus the ionic activity ratio (Mg2+)/(Ca2+) in the solution equilibrated with the solid).
Results Table 1 shows the chemical composition of the biomineral samples before and after being deproteinated by the plasma ashing and then washed. Although the apparent contents of Ca and P of all the samples increased following the treatment (due to decomposition of organic matter), their molar ratios did not change significantly, indicating that the mineral phase was unaltered by the treatment. This notion was supported by XRD and FTIR analyses of the treated solid samples (data not shown). As to the total Mg content of each of the tissues, the present analyses confirmed results reported in the literature, showing that the Mg content of human enamel (0.22 wt%) was lower than that (0.90 wt%) of human dentin. Interestingly, the Mg content (0.73 wt%) of younger porcine dentin (in unerupted teeth) was substantially lower than that of the erupted deciduous dentin of the same species (1.33 wt%) or that of human dentin in erupted teeth. The Mg content (0.52 wt%) of porcine bone was lower than that of the dentin samples in the same species. Chemical analyses of the water solutions used to wash the plasma-ashed samples revealed that only small amounts of Mg were released from the solid. The summed amounts of Mg solubilized during the repeated washing procedures amounted to less than 5% of the total Mg in all biomineral samples (see Table 4). We also analyzed the composition of some of the equilibrated solid samples, but no appreciable changes in the total Ca, P, and CO3 contents were detected (data not shown). Table 2 shows typical analytical results of the ionic composition of experimental solutions before and after equilibration with hydroxyapatite crystals. Following equilibration, the Mg concentrations decreased consistently in all systems, whereas both Ca and P concentrations increased. The pH values of the equilibrium solutions were in narrow ranges (6.2-6.5). It is pertinent to note that the equilibrium solutions were undersaturated with respect to Ca phosphate minerals such as dicalcium phosphate dihydrate, octacalcium phosphate, and hydroxyapatite. Consequently, precipitation of those Ca phosphates was precluded during equilibration. The increments of the Ca in solution were substantially higher (5-22 times) than the accompanying increments of phosphorus, which cannot be explained by a simple dissolution of the original crystals having a Ca/P molar ratio of 1.55. Examination of the equilibrated solid samples by XRD, FTIR, and EM did not provide evidence for substantial dis-
T. Aoba et al.: Adsorption of Mg z+ onto Apatite Crystals
146
Table 1. Chemical composition of biominerals before and after plasma ashing
Sample a
Chemical content
Molar ratio
Weightb loss wt%
Ca wt%
P wt%
Mg wt%
C03 wt%
CaJP
Mg/Ca
Human enamel
A B
-
Calcified Tissue International 9 1992 Springer-Verlag New York Inc.
Competitive Adsorption of Magnesium and Calcium Ions onto Synthetic and Biological Apatites T. Aoba, 1 E. C. Moreno, 1 and S. Shimoda 2 1Forsyth Dental Center, 140 Fenway, Boston, MA 02115; and 2Tsurumi University School of Dentistry, Yokohama 230, Japan Received September 16, 1991, and in revised form November 4, 1991
Summary. Magnesium (Mg) is a conspicuous constituent of hard tissues but its possible role in biomineralization is poorly understood. It is possible that Mg 2+ adsorbed onto bioapatites may contribute to the modulation of crystal growth as such inhibitory activity has been reported for synthetic apatites. The present study was undertaken to determine the adsorption isotherms of Mg ions onto synthetic apatites and biominerals in tooth and bone tissues in the presence of other ions of natural occurrence. Synthetic crystals used as adsorbents were hydroxyapatite and, as a better prototype for the biomineral, Mg-containing carbonatoapatite. Human enamel and dentin materials were obtained from extracted, caries-free, permanent teeth. Porcine dentin materials at two developmental stages were obtained from erupted deciduous and unerupted permanent teeth of a 6-month-old slaughtered piglet. Porcine bone was obtained from the cortical portion of the mandible of the same animal. All biomineral samples were pulverized and then treated by plasma ashing (deproteination) at about 60~ Each of the powdered samples was equilibrated in solutions containing various initial concentrations of Mg 2+, Ca 2+, and Na § (or K +) as nitrate salts. Following equilibration, concentrations (and activities) of magnesium and calcium ions in the experimental solution were determined. The pH values of the equilibrium solutions were in the range of 6.2-6.5. Experimental data of the Mg adsorption onto hydroxyapatite were interpreted on the basis of a Langmuir-type model for binary systems assuming competition of Mg 2§ and Ca 2 + for the same adsorption sites on the crystal surfaces of the apatites. According to this model, the adsorbed Mg is expressed as a function of the ionic activity ratio (Mg 2 +)/(Ca 2 § in the equilibrium solution. The model contains two parameters, the adsorption selectivity constant Ks and the maximum number of adsorption sites N (ixmol/g). The numerical values of Ks were similar for all adsorbents used (synthetic and biological) and indicated the preferential adsorption of Ca 2 § probably due to spacial restrictions extending to the very surface of the crystals. The initial level of Mg 2 § in the surface pool was different in the various biominerals, probably reflecting the composition of fluid in which the biominerals were formed. Whereas the surface pool of Mg of human enamel was marginal, only 5% of the total Mg, significant fractions of the total Mg in human and porcine dentins (about 2030%), and porcine bone (about 40%) existed on the crystal surfaces. There were significant differences in the total Mg and the value of the parameter N between young (unerupted) and mature (erupted) dentin minerals. It was ascertained that
Offprint requests to: T. Aoba
the occupancy of adsorption sites by Mg ions became greater with maturation of the dentin tissues. The overall results suggest that the Mg-mineral interaction in tooth and bone tissues may be a highly tissue-specific process, presumably reflecting differences in fluid composition (particularly Ca and Mg activities) responsible for biomineralization. Key words: Adsorption - Magnesium - Calcium - Apatite crystals - Enamel - Dentin - Bone.
Various reports [1-3] suggest that pyrophosphate (PPi), citrate, and magnesium (Mg) are potent inhibitors of calcium phosphate precipitation in biological fluids, such as plasma ultrafiltrate and urine. It was also established that, for equimolar concentrations, PPi is the most effective inhibitor, followed by citrate and magnesium in decreasing order of effectiveness [1]. However, when the inhibitors are used at physiological concentrations, Mg (e.g., about 0.8 mM in plasma ultrafiltrate) provided the most significant inhibition of seeded crystal growth of hydroxyapatite [1, 4, 5]. Most likely, the inhibitory activity is related to the adsorption of these ions onto growth sites on the surface of the apatite crystals. Indeed, we reported [6] that the inhibitory activity of PPi on the apatite crystal growth can be explained on the basis of its adsorption onto the seed crystals, which is described by the Hill-de Boer adsorption model [7]. However, to our best knowledge, no distinct model for the adsorption of Mg onto apatite crystals has been reported. As expected from the ubiquitous existence of Mg in various hard tissues, fluids separated from mineralizing zones have significant concentrations of Mg ions. Of particular interest are the differences in Mg concentration observed in different tissues. Thus, the cartilage fluid separated from chicken hypertrophic zone [8] and the enamel fluid from developing enamel tissue [9] have Mg concentrations of about 0.9 mM, similar to the Mg concentration that provides an effective inhibition of calcium phosphate precipitation [1, 4, 5], whereas the Mg concentration in predentin fluid [10] is much lower, about 0.2 mM. Although the differences observed in Mg concentration in the fluids of calcifying tissues suggest differences in the roles played by Mg ions in biomineralization from tissue to tissue, the limited knowledge on this subject precludes the assignment of specific function(s) for Mg in the mineralization of enamel and cartilage. Furthermore, in spite of the abundant information on the total Mg contents in tooth and bone tissues, our knowledge is still limited as to its status, whether it is present as a free ion or organic-bound in the fluid phase, adsorbed onto crystal surfaces, or incorporated into the crystal lattice of the
144 mineral. Q u a n t i t a t i v e i n f o r m a t i o n as to t h e labile (in solution), surface, a n d stable pools o f M g ions in t o o t h a n d b o n e tissues is e s s e n t i a l to (1) a s s e s s p o s s i b l e f u n c t i o n a l c o n t r i b u tions o f M g ions in t h e r e g u l a t i o n o f b i o m i n e r a l i z a t i o n a n d (2) d e t e r m i n e t h e n a t u r e a n d p r o p e r t i e s (e.g., crystalline lattice s t o i c h i o m e t r y a n d solubility) o f the f o r m e d b i o m i n e r a l . Furt h e r m o r e , it is p o s s i b l e t h a t t h e r e p o r t e d [3-5] c r y s t a l g r o w t h inhibition b y M g 2+ as well as its stabilization of m e t a s t a b l e calcium p h o s p h a t e s [11, 12] is r e l a t e d to a d s o r p t i o n of the ion o n t o specific sites o f t h e solid p h a s e . A similar role m a y b e p l a y e d b y the M g 2+ o n crystal s u r f a c e s in b i o m i n e r a l i z a tion p r o c e s s e s . T h e first step in t h e t e s t o f this h y p o t h e s i s , u n d e r t a k e n in this p u b l i c a t i o n , is to d e t e r m i n e t h e a d s o r p t i o n o f M g 2+ o n t o well-defined, s y n t h e t i c apatitic crystals, particularly in t h e p r e s e n c e o f C a 2 + ions as t h e l a t t e r are ubiquitous in m i n e r a l i z a t i o n z o n e s . W e also r e p o r t a m o d e l for the a d s o r p t i o n o f M g z+ o n t o d e n t a l e n a m e l , d e n t i n , a n d b o n e minerals. Finally, w e outline a p r o c e d u r e t h a t , using simple a d s o r p t i o n m o d e l s , p e r m i t s t h e d e t e r m i n a t i o n of the surface p o o l o f M g 2+ in t h e b i o a p a t i t e s .
Materials and Methods
Synthetic Apatites, Enamel, Dentin, and Bone Synthetic crystalline apatite samples used were hydroxyapatite, CaOHAp, having a composition of Ca 37.5%, total P 18.7%, Mg 0.00%, and CO 3 0.0%. A better prototype of biomineral was a magnesiumcontaining carbonatoapatite, (Ca,Mg)-CO3Ap, prepared in our laboratories, with a composition of Ca 35.2%, total P 17.7%, Mg 0.39%, and CO3 3.0% [13]. Specific surface areas of OH-Ap and (Ca,Mg)CO3AP, determined by N 2 adsorption (Quantachrome, NY), were 26.5 and 6.7 m2/g, respectively. Biomineral samples used were human enamel and dentin, and porcine dentin and bone. Sampling procedures of human enamel and dentin were reported previously in detail [14]. Briefly, human, caries-free permanent premolars and molars, extracted for orthodontic or periodontal reasons, were collected and stored in deionized water containing 0.5% (v/v) chloroform at 4~ Prior to preparation of the enamel samples, the enamel surface was thoroughly cleaned with pumice powder and a rotary rubber cup. The underlying dentin was ground off under running water; the grinding was advanced well beyond the dentin-enamel junction. For collection of dentin tissue, sound roots of the teeth were separated and then enough exterior surface (as well as interior surface facing the pulp cavity) was removed using a dental burr under water. The prepared hollow crown (enamel) and dentin specimens were placed separately in a sonicator with water (to remove fine debris) and dried at room temperature. Porcine materials were obtained from a 6-month-old slaughtered piglet. Dentin was dissected from erupted deciduous and unerupted permanent teeth; in this study, no special care was given to possible local differences in Mg content within enamel or dentin (apical or incisal aspects). Dentin tissue from each of the tooth groups was cut in the form of blocks with a dental discoid blade, cooling with running water. The dissected dentin blocks were treated in a similar way described for the human sample. For preparation of the porcine bone sample, the cortical portion of the mandible was cut with a circular saw, cooling with running water. The dissected bone block was transferred into liquid nitrogen and then crushed into small pieces while frozen. Each of the enamel, dentin, and bone samples was pulverized with an agate mortar and pestle. The resulting powder was passed through a 270 mesh sieve and stored at - 20~ until used. Prior to the adsorption experiments, each pooled sample was ashed in a plasma (oxygen activated by a radio frequency) at about 60~ as reported previously [15]. The weight loss of the sample was monitored and the treatment was continued until a constant weight was obtained, which usually occurred within 48 hours. Following the treatment, the ashed sample (about 100 mg) was placed in a centrifuge tube. In order to wash out debris of decomposed organic
T. Aoba et al.: Adsorption of Mg z+ onto Apatite Crystals matter, one ml of deionized water, precooled on ice, was added to the tube, and the resulting slurry was dispersed with a sonicator. The suspension was separated by centrifugation at 10,000 x g for 10 minutes. The supernatant was carefully removed, and the sediment was redispersed with the addition of another 1 ml of water. The same procedure was repeated three times. Following the last washing, the sediment was freeze-dried and then used as required in the adsorption experiments. Volumes of the recovered supernatants were determined by weighing (assuming unitary density), and aliquots of the supernatants were used for analysis of Mg ions displaced from the solid in the solution (see below).
Adsorption of Mg e + onto Synthetic Apatites and Biominerals ACS-grade reagents Mg(NO3)26H20, Ca(NO3)24HzO, NaNO3, and KNO3 (Aldrich Chem Co., Milwaukee, WI) were used to prepare stock solutions. In the systems in which the adsorbents were synthetic apatites, the initial concentrations of magnesium and calcium in the experimental solutions were varied in the range of 0.0-10 raM. A total of 19 systems were used with Ca-OHAp and 15 with MgCO3Ap. Most of the adsorption experiments were conducted in the presence of 160 mM of either NaNO3 or KNO3 to provide an ionic strength similar to that found in physiological fluids. Some of the experiments (seven systems) were conducted at various concentrations of Na + ions (0, 50, and 100 mM), in order to investigate the effect of monovalent cations on the Mg adsorption onto the crystals. Prior to each experimentation, experimental solutions (50-100 ml total volume) having specified concentrations of total Mg and Ca were prepared from the stock solutions and then pure nitrogen was bubbled through them to expel CO2. Accurately weighed amounts of the crystals (in most cases, 10 mg) were added to 1 ml of the experimental solution. The resulting slurry was mixed initially with the aid of an ultrasonicator and then with a vortex mixer, followed by endover-over rotation at 25(-+ 1)~ In preliminary studies, with 10-100 mg of synthetic crystals added to 1 ml of the experimental solutions, Mg concentrations reached plateau values within 1 hour and did not change further over a 24-hour period. Therefore, the solid-solution equilibration in the following experiments was conducted for 2 hours unless otherwise specified. After equilibration, the suspension was centrifuged at 10,000 z g for 10 minutes. The recovered supernatant, as well as an aliquot of the original experimental solution, was used for chemical analyses (see below). The sedimented solid was freeze-dried and then used for analyses by X-ray diffraction (XRD), Fourier-transform infrared spectroscopy (FTIR), and electron microscopy (EM). Procedures for the instrumental analyses were reported previously in detail [13, 15]. In studies with each of the biomineral samples, a weighed powder sample (10 rag) was equilibrated in 1 ml of the experimental solution containing 160 mM NaNO 3. Initial concentrations of the total Mg and Ca were adjusted in the ranges of 0-3 mM and 0-1 mM, respectively. The rest of the experimental procedures were the same as described above. Eight systems were used with each of the human enamel, human dentin, young porcine dentin, and bone minerals; seven systems were used with older porcine dentin as the adsorbents.
Chemical Analysis The total Mg and Ca concentrations were measured by atomic absorption (AA) spectrophotometry. Sodium and potassium were determined with the same instrument in the emission mode. Standard solutions for Ca determination were prepared with CaCO3 (dissolved in dilute HCI solution); standard solutions for AA analyses of Mg, Na, and K were obtained commercially (Fisher Sci, Pittsburgh, PA). The total phosphorus was determined colorimetrically [16]. The estimated errors of these chemical determinations were as follows: Mg and Ca 3%, total P 1.5%, and Na and K 4%. The pH value of the experimental solution was determined using a glass-reference combination electrode (Fisher) and a pH meter (Orion, model 901, Cambridge, MA). Carbonate concentrations of the supernatant sam-
T. Aoba et al.: Adsorption of Mg2+ onto Apatite Crystals
145
pies were determined by ion chromatography (Dionex, Sunnyvale, CA) according to the method of Kreling and DeZwaan [17]. Based on the analytical data, the Mgz+ and Caz+ activities in the equilibrium solution were calculated using a generalized program, previously reported [18], that takes into account the formation of Mg and Ca ion-pairs with various forms of phosphate and carbonate ions.
Mg in the solution. Thus, the quantity ~Mg may be positive or negative (desorption of Mg2+), depending on the experimental conditions. Thus, equation (5) is rewritten as follows: R/(Q~ + 8Mg) = R/N + 1/K~N Rearranging the above equation, one obtains Qi § ~Mg = NKsR/(1 + KsR)
Adsorption Model of Mg Ions onto Apatite Crystal Surfaces Preliminary experiments indicated that adsorption of Mg was affected by the presence of Ca ions in the equilibrating solution and suggested a one-to-one exchange reaction on the apatitic surface. Consequently, the adsorption model adopted here considers a competing adsorption of Mg 2+ and Ca 2+ ions onto the apatite crystals (i.e., the two ions compete for the same adsorption sites). The equilibrium between ions in solution and on the crystal surface can then be expressed by the reaction C a ~ f ~ + Mg2+solution~-Mgsurface + CaZ+solution
(1)
Assuming that the exchangeable pool (i.e., the adsorption sites on the surface) accommodates a total of N moles of Ca2+ and Mg2+ per unit of surface area (or unit mass in the case of the biominerals), the equilibrium constant K for reaction (1) is given by K = QMg * (CaZ+)/[(N - QMg) * (Mg2+)]
(2)
in which parentheses represent activities of the ions enclosed and QMg stands for the moles of Mg2+ adsorbed per unit surface area or per unit of mass of adsorbent (the corresponding quantity for Ca2+ is given by N - QM~). Rearranging the above expression and making R = (Mg2+)/(Ca~+), the condition for equilibrium can be written as R/QMg = R/N + I/KN
(3)
This final equation is of the same form as that describing a Langmuir adsorption model, although the ratio R of Mgz+ and Caz+ activities is used here, instead of the Mg2+ activity in the equilibrated solution. The same equation can be derived by the use of a Langmuir adsorption model for a binary system [19] Qr~g = NKMg(MgZ+)/[1 + KMg(Mg2+) + Kca(Ca2+)] Qca = NKca(Ca2+)/[1 + KMg(Mg2+) + Kca(Ca2+)] N = QMg + Qca
(4)
in which KMg and Kc~ are the adsorption affinity constants of Mg2+ and Caz+ ions, respectively, onto the crystal surfaces. Calling Ks = KMg/Kc~ (the selectivity constant), combination of the three equations in (4) yields R/QMg = R/N + 1/KsN
(5)
which is formalistically identical to equation (3). If the model described by either equation (3) or (5) is correct, a plot of the experimental data in terms of R/QMg v e r s u s R should yield a linear relationship, and the numerical values of Ks and N can be obtained from the slope and intercept. Experimentally, numerical values of R were obtained based on the solution composition at the end of equilibration, whereas the corresponding QMg values were obtained by the difference in the total concentration ofMg 2+ ions in the experimental solutions before and after equilibration. In systems using Mg-containing apatite or the biominerals as adsorbents, it is reasonable to assume that part of the total Mg in the solid sample exists initially on the crystal surfaces in the same fashion as part of the Ca in Ca-OHAp is on the surface and exhibits exchange properties. Consequently, the above adsorption model should be modified to accommodate the initial surface pool of Mg ions, Qi. Calling gMg the adsorbed Mg experimentally determined after a given equilibration, the quantity Q~g (i.e., the total amount of Mgz+ in the adsorbed state on the crystals at equilibrium) is given by the sum of Qi and gMg. There is no a priori knowledge of the quantity Qi; also, the equilibrating solution may become enriched or impoverished in Mg2+ , depending on the ionic activities of Ca and
(6)
(7)
In equation (7), the three parameters N, Ks, and Qi are unknown. Their numerical values were obtained by the use of a nonlinear least-squares procedure together with the experimental data obtained from all the systems having a common adsorbent. The values of the three parameters are expected to be different for different adsorbents. Of particular interest are the values of Qi, the initial Mg present in the surface pool, and the value of Ks which, to some extent, should reflect the discrimination of Mg from the structure of Ca-apatites. In this publication, the term "adsorption isotherm" refers to the plot of the total amount of Mg in the surface pool Q~g (Qi plus the experimental quantity ~Mg) versus R (i.e., versus the ionic activity ratio (Mg2+)/(Ca2+) in the solution equilibrated with the solid).
Results Table 1 shows the chemical composition of the biomineral samples before and after being deproteinated by the plasma ashing and then washed. Although the apparent contents of Ca and P of all the samples increased following the treatment (due to decomposition of organic matter), their molar ratios did not change significantly, indicating that the mineral phase was unaltered by the treatment. This notion was supported by XRD and FTIR analyses of the treated solid samples (data not shown). As to the total Mg content of each of the tissues, the present analyses confirmed results reported in the literature, showing that the Mg content of human enamel (0.22 wt%) was lower than that (0.90 wt%) of human dentin. Interestingly, the Mg content (0.73 wt%) of younger porcine dentin (in unerupted teeth) was substantially lower than that of the erupted deciduous dentin of the same species (1.33 wt%) or that of human dentin in erupted teeth. The Mg content (0.52 wt%) of porcine bone was lower than that of the dentin samples in the same species. Chemical analyses of the water solutions used to wash the plasma-ashed samples revealed that only small amounts of Mg were released from the solid. The summed amounts of Mg solubilized during the repeated washing procedures amounted to less than 5% of the total Mg in all biomineral samples (see Table 4). We also analyzed the composition of some of the equilibrated solid samples, but no appreciable changes in the total Ca, P, and CO3 contents were detected (data not shown). Table 2 shows typical analytical results of the ionic composition of experimental solutions before and after equilibration with hydroxyapatite crystals. Following equilibration, the Mg concentrations decreased consistently in all systems, whereas both Ca and P concentrations increased. The pH values of the equilibrium solutions were in narrow ranges (6.2-6.5). It is pertinent to note that the equilibrium solutions were undersaturated with respect to Ca phosphate minerals such as dicalcium phosphate dihydrate, octacalcium phosphate, and hydroxyapatite. Consequently, precipitation of those Ca phosphates was precluded during equilibration. The increments of the Ca in solution were substantially higher (5-22 times) than the accompanying increments of phosphorus, which cannot be explained by a simple dissolution of the original crystals having a Ca/P molar ratio of 1.55. Examination of the equilibrated solid samples by XRD, FTIR, and EM did not provide evidence for substantial dis-
T. Aoba et al.: Adsorption of Mg z+ onto Apatite Crystals
146
Table 1. Chemical composition of biominerals before and after plasma ashing
Sample a
Chemical content
Molar ratio
Weightb loss wt%
Ca wt%
P wt%
Mg wt%
C03 wt%
CaJP
Mg/Ca
Human enamel
A B
-
