- Email: [email protected]
Adsorption of immunogamma globulin onto various synthetic calcium hydroxyapatite particles
Journal of Colloid and Interface Science 273 (2004) 406–413 www.elsevier.com/locate/jcis
Adsorption of immunogamma globulin onto various synthetic calcium hydroxyapatite particles Kazuhiko Kandori,∗ Kaori Miyagawa, and Tatsuo Ishikawa School of Chemistry, Osaka University of Education, Asahigaoka 4-698-1, Kashiwara, Osaka 582-8582, Japan Received 25 April 2003; accepted 29 January 2004
Abstract This paper presents data on adsorption of immunogamma globulin (IgG) onto synthetic rodlike calcium hydroxyapatite particles (CaHaps) with various particle lengths and calcium/phosphate (Ca/P) atomic ratios ranging from 1.54 to 1.65 and compares the obtained results to those of acidic (bovine serum albumin, BSA), neutral (myoglobin, MGB), and basic (lysozyme, LSZ) proteins reported before. The effect of electrolyte concentration on IgG adsorption was also examined. The initial rate of IgG adsorption was similar to that of BSA and was slower than that of MGB and LSZ. This fact was interpreted by the difference in the structural stability and molecular weight of these proteins. The isotherms of IgG adsorption onto the CaHap particles were of pseudo-Langmuir type. The saturated amount of adsorbed IgG IgG values (ns ) for the particles with mean particle length less than 70 nm decreased with increasing Ca/P ratio. The adsorption behavior IgG of IgG molecules was very similar to that of basic LSZ, though IgG has zero net charge. The ns value was increased with increased mean particle length of CaHaps; the relationship was less significant than that for BSA but similar to those for MGB and LSZ. The similar adsorption behavior of IgG and LSZ suggested that the Fab parts of IgG molecules preferentially adsorb onto CaHap to provide the reversed Y-shaped conformation of IgG. The change of the adsorption mode of IgG molecules from the reversed Y-shaped conformation IgG to side-on by “spreading” the Fc part of IgG molecules onto the particle surface over a longer adsorption time was suggested. The ns value was increased with increasing electrolyte concentration by screening the intra- and intermolecular electrostatic interactions of proteins. 2004 Elsevier Inc. All rights reserved. Keywords: Protein adsorption; Immunogamma globulin; Synthetic calcium hydroxyapatite particles; Bovine serum albumin; Myoglobin; Lysozyme; Electrophoretic mobility
1. Introduction Adsorption of IgG onto solid surfaces is of great interest for biomedical applications, such as immunoassays and biosensors. IgG or antibodies are proteins synthesized by an animal in response to the presence of a foreign substance called an antigen [1–3]. These proteins, naturally designed with extraordinary specificity and binding affinity for a given antigen, are used as diagnostic reagents. Therefore, many investigations have been done using different adsorbents such as silica gel [4,5], polymer latex [6–8], and Teflon [9,10] particles. In the development of diagnostic reagents, the main goal is to achieve stable attachment of IgG molecules at the solid surfaces without disrupting or masking the biological * Corresponding author.
E-mail address: [email protected] (K. Kandori). 0021-9797/$ – see front matter 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.jcis.2004.01.069
functions of the protein, such as selectivity and specificity. An IgG molecule is composed of four polypeptide chains, two heavy and two light chains, and they are grouped into two domains [1,2,11]. The molecule is composed of two Fab segments and one Fc segment and forms a “Y-shaped” conformation. Because the antigen-binding sites are located on the far ends of the Fab segments, it is desired that the IgG molecules be adsorbed onto the adsorbent surface by anchoring Fc parts with exposing the Fab segments toward to the outside. On the other hand, it is well known that calcium hydroxyapatite [Ca10 (PO4 )6 (OH)2 , CaHap] is a major inorganic component of bone and tooth and possesses a high affinity for proteins. Since CaHap is in the space group P 63 /m and its unit cell dimensions are a = b = 0.943 nm and c = 0.688 nm, possessing two different binding sites [12,13], called C and P sites on the crystal surface, it has a multiple-
K. Kandori et al. / Journal of Colloid and Interface Science 273 (2004) 406–413
site binding character for proteins. After dispersed CaHap particles in aqueous media, calcium atoms (C sites) are exposed on the CaHap surface by dissolution of OH− ions at the particle surface. Therefore, the C sites are arranged on ac or bc crystal faces in a rectangular manner with the interdistance in the a or b directions equal to 0.943 nm and the interdistance in the c direction equal to 0.344 nm (c/2). The P sites are arranged hexagonally on the ab crystal face with a minimal distance of 0.943 nm. The C sites are rich in calcium ions or positive charges and thus bind to acidic groups of proteins, but the P sites lack calcium ions or positive charges and therefore attach to basic groups of proteins. The fundamental studies of the adsorption of acidic (BSA: isoelectric point (iep) = 4.7), neutral (MGB: iep = 7.0), and basic (LSZ: iep = 11.1) proteins onto various kinds of synthetic hydroxyapatites (Hap), such as calcium Hap (CaHap), strontium Hap (SrHap), calcium–strontium Hap (CaSrHap), and carbonate-containing CaHap (Cap), have been carried out by the authors [14–20]. We clarified in these studies that the saturated amount of adsorbed BSA, nBSA , depends on the molar ratio of s cations (Ca, Sr, and/or Ca + Sr) to phosphate (cation/P) in the materials used; the nBSA increased with increase in the s cation/P ratio and was explained by the electrostatic attractive forces between negatively charged BSA and the less negatively charged CaHaps with high cation/P ratios at pH 6. The same result was further observed for the adsorption of BSA onto CaHap [14], CaSrHap [16], and Cap [17] particles. In contrast, however, the saturated amounts of adsorbed LSZ (nLSZ s ) on SrHap and CaSrHap particles decreased with increased cation/P ratio and approach zero around at cation/P = 1.70 [19], whereas various Haps exhibited constant values of nMGB and no remarkable relationship with s cation/P of the particles was detected for the neutral protein MGB [20]. The adsorption of IgG onto CaHap has been reported so far by two dental research groups. Rölla et al. [21] reported that immunoglobulins, IgA and IgG, constitute a part of the pellicle on the tooth surface. Furthermore, Sasaki et al. [22] and Eggen and Rölla [23] disclosed that the effect of NaF surface treatment of CaHap on the adsorption of IgG. To our knowledge, however, no quantitative data have been reported in the literature on the adsorption of IgG. This paper presents adsorption data of IgG onto synthetic CaHaps with various sizes and Ca/P ratios ranging from 1.54 to 1.65 and compares the results obtained to those of acidic, neutral, and basic proteins reported before. Comparison studies of these systems may contribute to further understanding of the mechanism of IgG adsorption onto CaHaps. The effect of electrolyte concentration on the IgG adsorption was also examined. The results also might be beneficial for applying CaHap to the HPLC, immunoassay, and biosensor systems.
407
Fig. 1. Electron micrographs of typical CaHap particles (Nos. 3, 6, 10 and 13) given in Table 1.
2. Experimental 2.1. Materials and methods Thirteen kinds of rodlike CaHap particles with different particle lengths ranging from 24 to 83 nm but similar diameter of 5–16 nm, as shown in Table 1, used in this study were synthesized by aging the precipitates at 100 ◦ C formed from the reaction of a solution of Ca(OH)2 with H2 PO4 in a Teflon vessel for 48 h under CO2 -free conditions. The particle length was controlled by changing the molar ratio of Ca(OH)2 to H3 PO4 in the solutions. The recipe for these particle preparations was reported in detail elsewhere [14,15]. These particles were characterized by conventional methods of XRD, TG-DTA, ICP-AES, TEM, N2 adsorption, and zeta potential measurement [14,15]. The mean particle size was obtained by measuring the length and diameter of 200 particles in each TEM micrograph. The adsorption isotherm of N2 was measured at liquid nitrogen temperatures using a computerized automatic volumetric apparatus built in house. Prior to the measurement, the samples were evacuated at 300 ◦ C for 2 h. The specific surface area was determined within a high accuracy of ±2% by fitting the adsorption isotherms to the BET equation. The typical TEM pictures of several samples are displayed in Fig. 1. It should be mentioned here again that the CaHap particles precipitated at higher concentrations of H3 PO4 (Nos. 10–13) are large rods with mean particle lengths over 70 nm. The polyclonal IgG (99% purity), fraction II and III, was purchased from Sigma Co. (G-5009). The electrophoretic mobility (EM) of the particles was measured using an electrophoresis apparatus as reported before.
408
K. Kandori et al. / Journal of Colloid and Interface Science 273 (2004) 406–413
Table 1 IgG Properties of various kinds of CaHap and saturated amounts of adsorbed IgG (ns ) IgG
No.
[H3 PO4 ] (mol)
Ca:PO4 atomic ratio
Specific surface area (m2 g−1 )
Crystallite size of (002) plane (nm)
Mean particle size (nm)
ns (mg m−2 )
1 2 3 4 5 6 7 8 9 10 11 12 13
0.24 0.24 0.24 0.24 0.24 0.235 0.24 0.24 0.247 0.249 0.252 0.265 0.25
1.54 1.56 1.58 1.59 1.60 1.62 1.62 1.64 1.58 1.58 1.59 1.61 1.65
96 91 90 101 92 94 104 104 90 97 97 112 101
46 46 40 41 23 34 38 16 29 27 17 23 24
46 × 7 36 × 8 53 × 9 69 × 8 35 × 16 33 × 7 24 × 12 29 × 13 60 × 8 83 × 7 76 × 5 83 × 7 77 × 5
0.50 0.43 0.46 0.28 0.27 0.35 0.18 0.16 0.40 0.50 0.60 0.61 0.52
2.2. Protein adsorption experiment The protein adsorption experiment was carried out in 10 cm3 Nalgen polypropylene (PP) centrifugation tubes to which varying concentrations of IgG in 1 × 10−4 mol dm−3 KCl solution of pH 6 were added to disperse the CaHaps (150 mg) by the batch method given elsewhere [14–17]. Preliminary experiments revealed that the solubility of the CaHap particles employed in the present study in KCl solution is quite low under this neutral pH condition. The calcium and phosphate ions dissolved from CaHap particles in a KCl solution, assayed by the ICP-AES method, were ca. 2.5 × 10−4 and 1.7 × 10−5 mol dm−3 , respectively. These ion concentrations are considered to be far too low for varying the solubility of IgG molecules and the electric double layer of CaHap particles. The 10 mg cm−3 IgG solutions were prepared by dissolving IgG in 1 × 10−4 mol dm−3 KCl solution. However, since IgG did not completely dissolve, the clear supernatant after the solution was centrifuged was used for the experiment. The IgG concentration in the supernatant was spectroscopically determined using an extinction of 1.45 cm3 mg−1 at 280 nm [1,4]. The initial concentrations of IgG in PP tubes were varied from 0 to 7.5 mg cm−3 . The PP centrifugation tubes were gently rotated end over end at 15 ◦ C for 48 h in a thermostat. To determine the adsorption isotherms of IgG in the static condition, the concentrations of IgG in the supernatant after centrifugation were determined by the microburet method using an UV adsorption band at 310 nm. The pHs of IgG solution before and after adsorption were measured by Toa pH-meter. In contrast, to compare the initial rates of IgG adsorption with other proteins examined before, the amounts of adsorbed IgG were measured at various incubation times from 2 to 48 h using CaHap No. 4 particles. Preliminary experiments revealed high reproducibility of the method (±2%) using the PP centrifugation tube and no protein adsorption onto the PP tube could be detected. The effect of electrolyte on the adsorption behavior of IgG was also examined by varying the KCl concentration between 1.0 × 10−4 and 1.0 mol dm−3 .
Fig. 2. Kinetics of adsorption of IgG (2), MGB (E), BSA (!), and LSZ (P) from each single-protein system onto CaHap (No. 4).
3. Results and discussion 3.1. Initial rate of IgG adsorption onto CaHap (kinetics study) The initial rate of IgG adsorption onto CaHap was examined at an initial concentration of 7.0 mg cm−3 using CaHap No. 4 as displayed in Fig. 2. For the comparison, the data of BSA, LSZ, and MGB adsorption onto the same material from each single protein system as reported in our previous studies [20] are also shown in Fig. 2 (open symbols). It is obvious that the initial rates of IgG (2) and BSA (!) adsorption are lower than those of MGB (E) and LSZ (P). The kinetic data can usually be analyzed by a diffusion-control adsorption model under quiescent conditions; protein molecules reach the particle surface by Brownian motion. Under such conditions, the rate of arrival of the protein at the particle surface is given by the Ward and Tordai equation [25], J = Cp (D/πt)1/2 , where J is the flux of the protein per unit area of particle surface, Cp is the concentration of protein, t is the time, D is the diffusion coefficient of IgG in the solution (4.0 ×10−11 m2 s−1 ), and π is 3.14. Many experimental data
K. Kandori et al. / Journal of Colloid and Interface Science 273 (2004) 406–413
confirmed diffusion-controlled adsorption from nonflowing solutions at low concentration [26,27]. However, since the solutions dispersing CaHap in PP centrifugation tubes are gently rotated end over end in a thermostat and the concentration of IgG (7.0 mg cm−3 ) is high in the present study, the diffusion-control model can not be applied to our data. On the contrary, it is well known that the proteins having low structural stability, called “soft proteins” such as IgG, BSA, and MGB, tend to change their structure upon adsorption. This structural rearrangement of IgG and BSA with high molecular weight (IgG: 146,000 Da, BSA: 67,200 Da) upon adsorption, therefore, may be the reason for the slow initial adsorption rate as compared to LSZ (14,000 Da), with high structural stability, and MGB, with low structural stability but with low molecular weight (17,800 Da). This structural change of adsorbed IgG molecules will be discussed later. Another finding in Fig. 2 is that the saturated amount of adIgG ) sorbed IgG (ns ) is much less than those of BSA (nBSA s MGB and MGB (ns ). This result indicates that the affinity of IgG to the CaHap surface is low. As we have reported before, BSA molecules are strongly adsorbed onto the CaHap, mainly through a specific electrostatic attractive force between negatively charged carboxylic groups on BSA and positively charged C sites localized on CaHap surfaces, even though the surface net charge of CaHap is negative at pH 6 [14,15]. Therefore, this specific . On electrostatic adsorption process provides a high nBSA s the other hand, since MGB molecules are expected to have a slight positive net charge at pH 6, similar specific electrostatic interaction did not operate between positively charged amino groups of MGB and C sites. The authors therefore conclude in the study that neutral MGB molecules approach the negatively charged CaHap surface much more readily than BSA ones without such specific electrostatic adsorption. In addition, since it is also well known that MGB molecules are characterized as “soft” proteins with lower structure stability, similar to BSA, we presumed that MGB could be adsorbed onto both hydrophobic and hydrophilic surfaces under attractive and repulsive electrostatic conditions, similarly to BSA. Actually, Kondo and Mihara emphasized the large conformational change of MGB upon adsorption on the colloidal particles [24]. The same mechanism can be anticipated for IgG molecules, because the IgG molecules are expected to have a slight positive net charge at pH 6, the same as MGB, as will be shown in Figs. 3 and 4 in the next section. The detailed mechanism of IgG adsorption onto CaHap surface will hereinafter be discussed. 3.2. Adsorption of IgG onto various kinds of CaHaps (static study) Figs. 3 and 4 display the characteristic adsorption isotherms of IgG and their mobilities and the pH values of the IgG solution before and after adsorption for relatively small and large CaHap particles, respectively. To avoid confusing the data points, the representative data of EM and pH for Ca-
409
Fig. 3. Isotherms of IgG adsorption onto CaHap (Nos. 1–9) (a), EM of No. 2 (b), and solution pH of No. 5 (c).
Hap Nos. 2, 5, and 13 are shown in the figures. Actually, similar data were obtained for the other CaHap particles. Here, to simplify the figure, the pH values of IgG solution before adsorption (!) are plotted at the equilibrium concentration of the corresponding solution after adsorption ("). The isotherms of IgG adsorption onto these particles are of the pseudo-Langmuir type, as we have already reported for BSA, MGB, and LSZ; the adsorption isotherms steeply rose in the lower concentration region and attained saturation [15–21]. It can be seen in Figs. 3 and 4 that the pH values of the IgG solution without CaHap particles are around 5.0–6.2, but they are increased to ca. 6.5–7.0 after adsorption onto the CaHap particle. This pH rise is due to generation of OH− ions from the CaHap particle surface by dissolution, because CaHap particles are stable under the present conditions, i.e., only produce very low contents of calcium and phosphate ions, as described before. No phosphate buffer solution was used in this study. This result therefore strongly supports the mechanism of C sites formation on the ac or bc
410
K. Kandori et al. / Journal of Colloid and Interface Science 273 (2004) 406–413
IgG
Fig. 5. Theoretical ns
values for different adsorption configurations.
Fig. 4. Isotherms of IgG adsorption onto CaHap (Nos. 10–13) (a), EM of No. 13 (b), and solution pH of No. 13 (c).
crystal surfaces was described in Section 1. The EM value shows nearly zero over the IgG concentration investigated, but the sign of the EM reverses from negative to positive with the IgG adsorption. This result indicates that the iep of IgG lies between pH 6 and 7. This fact is in agreement with the iep values of IgG reported (5.8–6.9) in other literatures [1,28,29]. The extremely small EM of CaHap particles suggests that IgG molecules can be regarded as neutral proteins under the conditions of the present study. IgG The plateau amounts of adsorbed IgG (ns ), corresponding to the saturated amounts of adsorbed IgG, are listed in Table 1. The IgG molecule has a Y-shaped conformation with dimensions of 7.0 × 6.3 × 3.1 nm3 for the Fc fragment and 8.25 × 5.0 × 3.8 nm3 for each of the two Fab fragments [2,9]. From this molecular size, we can evaluIgG ate the theoretical ns values for different adsorption configurations, i.e., end-on and side-on, by assuming that IgG molecules are adsorbed closely onto a flat surface (Fig. 5). IgG The experimental ns values obtained in the present study IgG (0.1–0.6 mg m−2 ) are less than all the theoretical ns val-
IgG
Fig. 6. (a) Plots of ns as a function of Ca/P ratio and (b) plots of IgG ns , nBSA , nMGB , and nLSZ as a function of particle length of CaHaps s s s (" > 70 nm, ! < 70 nm).
ues evaluated. Therefore, this result does not directly provide insight about the adsorption mode of IgG on CaHap. To discuss the adsorption of IgG onto CaHap in further IgG detail, we plotted the ns as a function of the Ca/P ratios of CaHap particles because our previous studies revealed that the adsorption amounts of proteins, especially for acidic and basic proteins, are strongly dependent on the Ca/P raIgG tio [19]. Fig. 6a depicts the change in the ns values with the Ca/P ratio. The data for CaHap particles with relatively large mean particle lengths over 70 nm are plotted by filled
K. Kandori et al. / Journal of Colloid and Interface Science 273 (2004) 406–413 IgG
circles. It is noteworthy that the ns values for the particles with mean particle lengths less than 70 nm (!) decrease with increased Ca/P ratio, though those for the particles over 70 nm in the mean particle length are almost constant around 0.5–0.6 mg m−2 . The authors previously found the important effect of electrostatic attractive force between protein and CaHap [18]. Since the large fraction of positively charged C sites is advantageous for the adsorption of negwas increased atively charged acidic BSA molecules, nBSA s with increased Ca/P ratios. In contrast, LSZ of basic proteins was decreased with exhibited an opposite relationship; nLSZ s increased Ca/P ratio and finally reached almost zero around Ca/P = 1.70. Unlike BSA and LSZ, no relationship between and the Ca/P value could be detected for neutral MGB. nMGB s The results observed for IgG molecules in Fig. 6a are very similar to those for basic LSZ reported before, though IgG has only slight positive charge. This unexpected but attractive result strongly suggests that the mode of IgG molecules adsorption onto CaHap is similar to that of basic LSZ ones but not to that of neutral MGB ones. IgG In contrast, the ns values for the relatively large CaHap particles (") do not follow this relationship. The authors reported the importance of C sites for the adsorption of acidic depends strongly protein BSA in the previous paper; nBSA s on the particle length, more markedly than on the Ca/P ratios, by the specific electrostatic interaction between negatively charged carboxylic acid groups of BSA and positively charged C sites on the exposed ac or bc crystal surfaces [18]. To make clear the behavior of IgG adsorption onto CaHaps, IgG therefore, we plotted the ns values as a function of the mean particle length for all the CaHap particles employed in Fig. 6b together with the results of BSA, MGB, and LSZ reported before by dotted lines [20]. It is easy to recognize IgG that the ns value is increased with increased mean particle length of CaHaps, the relationship is less significant than for BSA but similar to those for MGB and LSZ. Since IgG has a slight positive net charge as well as MGB, similar adsorption behavior can be expected on both proteins. But the slope of the line is close to LSZ, indicating that the IgG ns values are much less correlated with the C site of CaHap. This fact agrees with the results in Fig. 6a, suggesting that the neutral IgG behaves as a basic protein with high structural stability. Norde et al. reported that the adsorption behavior of the Fc domain is typically that of a “soft protein;” i.e., it adsorbs onto a hydrophilic, electrostatically repelling surface, whereas the “hard” Fab parts only adsorb if the hydrophilic silica under them is electrostatically attracted [30]. According to this result, it can be presumed that Fab parts of IgG molecules preferentially adsorb onto CaHap to take a reversed Y-shaped conformation (Fig. 5b), because IgG molecules and CaHap surfaces are electrically attracted under the conditions examined. The configuration of IgG molecules may further support this mechanism as follows: the Fab fragment of the IgG molecule is NH2 -terminal to give a positive charge, though the Fc one is COOH-terminal to provided negative charge [1,4], indicating a higher electro-
411
static attractive force between the Fab fragment and CaHap with negative net-charge. In the literature, it is inferred that IgG adsorption preferentially takes place with the Fc domain because this part is more hydrophobic and flexible than Fab part [2,8,31,32]. However, it can be assumed from the result in Fig. 6 that this mechanism cannot be applied to the adsorption of IgG onto hydrophilic and negatively charged CaHap particles at neutral pH conditions. It has also reported that adsorbed IgG molecules become more tightly bound as the contact time with the adsorbent surface increases [2,28,29]. This observation is due to slow structural rearrangements within the IgG molecules in order to optimize interactions with the adsorbent surface [1]. According to this concept, it is concluded that the conformational change is favorable for IgG, especially for the flexible Fc part. The conformational change, called “structural relaxation,” implies optimization of the protein–particle surface interactions by developing a larger number of contacts between them. Therefore, it seems reasonable to conclude that the mode of IgG molecule adsorption is changed from the reversed Y-shaped conformation to side-on within a longer adsorption time; a certain degree of “spreading” of the Fc part of the IgG molecule may take place. 3.3. Effects of electrolytes on the adsorption of IgG onto CaHap It is well known that the interactions between proteins and surface are influenced by ionic strength in the system [8,33–35]. Therefore, we examined the effects of electrolyte on the IgG adsorption onto CaHap (No. 12) by changing the concentration of KCl from 1.0×10−4 to 1.0 mol dm−3 . Prior to this experiment, we also checked the solubility of CaHap particles in these KCl solutions. However, CaHap particles exhibited high stability at all the KCl concentrations employed in this study, as described in a previous section, and no differences in calcium and phosphate concentrations and solution pH were observed under these experimental conditions. This result implies that the variation of adsorption of IgG onto CaHap is attributed to the effect of electrolytes added. The obtained adsorption isotherms measured at varied KCl concentrations are displayed in Fig. 7. There is a considerable increase in the adsorption as the KCl concentration is IgG increased up to 1.0×10−1 mol dm−3 (1), but ns is slightly −1 −3 decreased at 5.0 × 10 mol dm (filled symbols); the IgG ns exhibits a maximum at 1.0 × 10−1 mol dm−3 . Suzawa and Murakami explained the increase of nBSA onto polymer s latex with increasing electrolyte concentration by three reasons [8]: (1) The electrostatic repulsions in the interiors of protein molecules are neutralized by the addition of electrolytes, and this favors a compact protein structure. Namely, the conformational stability of protein molecules increases with increasing the ionic strength [36,37]. (2) The lateral repulsion between adsorbed protein molecules is weakened
412
K. Kandori et al. / Journal of Colloid and Interface Science 273 (2004) 406–413
Acknowledgments The authors thank Masao Fukusumi, Osaka Municipal Technical Research Institute, for help with the TEM. This work has been supported in part by a Grant-in-Aid for Scientific Research (B) from the Ministry of Education, Science, Sports and Culture.
References
Fig. 7. Isotherms of IgG adsorption onto CaHap (No. 12) at various KCl concentrations: KCI concentration: (e) 1.0 × 10−4 , (P) 1.0 × 10−3 , (!) 1.0 × 10−2 , (1) 1.0 × 10−1 , (") 5.0 × 10−1 , and (Q) 1.0 mol dm−3 .
with increase of ionic strength; thus more molecules can adsorb onto the given surface area. (3) The electrostatic repulsion between protein molecules and particle surface weakens, so protein molecules can more easily adsorb. Since IgG and CaHap are oppositely charged, reason (3) is unsuitable for explaining the present result. The appearance of maximum adsorption at 1.0 × 10−1 mol dm−3 indicates that there is an optimum KCl concentration for reasons (1) and (2). The secondary and tertiary structure of IgG could be changed at IgG higher KCl concentration and lead to depression of the ns . But the detailed mechanism of this fact is not clear at the moment and is left for subsequent papers.
4. Conclusion The initial rates of IgG and BSA adsorption are slower than those for MGB and LSZ. This fact is interpreted by the difference in the structural stability of these proteins. The IgG and BSA with less structural stability exhibit structural IgG rearrangement. The ns values for the particles with the short mean particle length are decreased with increased Ca/P ratio. This adsorption of IgG molecules is very similar to that of basic protein LSZ, though IgG has a slight positive net IgG charge as well as MGB. Therefore, the ns value is slightly increased with increased mean particle length of CaHaps, similarly to LSZ. The Fab parts of IgG molecules may preferentially adsorb onto CaHap to take the reversed Y-shaped conformation, but after a longer adsorption time the adsorption mode of IgG molecules could be changed to side-on by spreading Fc fragments of IgG molecules on the CaHap surface. The adsorption of IgG is enhanced by screening the intra- and intermolecular electrostatic interactions of protein molecules with addition of electrolyte.
[1] C.E. Giacomelli, in: A.T. Hubbard (Ed.), Encyclopedia of Surface and Colloid Science, Dekker, New York, 2002, p. 418. [2] J. Buijs, W. Norde, J.W.Th. Lichtenbelt, Langmuir 12 (1996) 1605. [3] M. Okubo, Y. Yamamoto, M. Uno, S. Kamei, T. Matsumoto, Colloid Polym. Sci. 265 (1987) 1061. [4] J.W.Th. Lichtenbelt, W.J.M. Henvelsland, M.E. Oldenzeel, R.L. Zsom, Colloids Surf. B Biointerfaces 1 (1993) 75. [5] A. Kamyshny, O. Toledano, S. Magdassi, Colloids Surf. B Biointerfaces 13 (1999) 187. [6] R. van Erp, Y.E.M. Linders, A.P.G. van Sommersen, T.C.J. Gribnau, J. Immunol. Methods 152 (1992) 191. [7] M. Okubo, Y. Yamamoto, M. Uno, S. Kamei, T. Matsumoto, Colloid Polym. Sci. 265 (1987) 1061. [8] T. Suzawa, T. Murakami, J. Colloid Interface Sci. 28 (1980) 266. [9] A.W.P. Vermeer, M.G.E.G. Bremer, W. Norde, Biochim. Biophys. Acta 1425 (1998) 1. [10] A.W.P. Vermeer, C.E. Giacomelli, W. Norde, Biochim. Biophys. Acta 1526 (2001) 61. [11] G.M. Edelman, in: J. Sterzl, I. Riha (Eds.), Development Aspects of Antibody Formation and Structure, Academic Press, New York, 1970, p. 381. [12] T. Kawasaki, M. Niikura, S. Takahashi, W. Kobayashi, Biochem. Int. 13 (1986) 969. [13] T. Kawasaki, K. Ikeda, S. Takahashi, Y. Kuboki, Eur. J. Biochem. 155 (1986) 249. [14] K. Kandori, S. Sawai, Y. Yamamoto, H. Saito, T. Ishikawa, Colloids Surf. 68 (1992) 283. [15] K. Kandori, Y. Yamamoto, H. Saito, T. Ishikawa, Colloids Surf. A 80 (1993) 287. [16] K. Kandori, M. Saito, H. Saito, A. Yasukawa, T. Ishikawa, Colloids Surf. A 94 (1995) 225. [17] K. Kandori, M. Saito, T. Takebe, A. Yasukawa, T. Ishikawa, J. Colloid Interface Sci. 174 (1995) 124. [18] K. Kandori, T. Shimizu, A. Yasukawa, T. Ishikawa, Colloids Surf. B 5 (1995) 81. [19] K. Kandori, N. Horigami, H. Kobayashi, A. Yasukawa, T. Ishikawa, J. Colloid Interface Sci. 191 (1997) 498. [20] K. Kandori, A. Fudo, T. Ishikawa, Phys. Chem. Chem. Phys. 2 (2000) 2015. [21] G. Rölla, J.E. Ciardi, W.B. Bowen, J. Dent. Res. 91 (1983) 186. [22] K. Sasaki, T. Mori, M. Terashita, M. Nagano, C. Uchiyama, Jpn. J. Oral Biol. 26 (1984) 1362. [23] H. Eggen, G. Rölla, J. Dent. Res. 91 (1983) 347. [24] A. Kondo, J. Mihara, J. Colloid Interface Sci. 177 (1996) 214. [25] A.F.D. Ward, L. Tordai, J. Chem. Phys. 14 (1946) 453. [26] G.A. Bornzin, I.F. Miller, J. Colloid Interface Sci. 86 (1982) 539. [27] P. Van Dulm, W. Norde, J. Colloid Interface Sci. 91 (1983) 248. [28] A. Martin, J. Puig, F. Galisteo, J. Serra, R. Hidalgo-Alvarez, J. Dispers. Sci. Technol. 13 (1992) 399. [29] J. Buijs, J.W.Th. Lichtenbelt, W. Norde, J. Lyklema, Colloids Surf. B Biointerfaces 5 (1995) 11. [30] W. Norde, in: J.M. Asna (Ed.), Polymeric Dispersion, Principles and Applications, Kluwer Academic, Dordrecht, 1997, p. 541.
K. Kandori et al. / Journal of Colloid and Interface Science 273 (2004) 406–413
[31] M.E. Soderquist, A.G. Walton, J. Colloid Interface Sci. 75 (1980) 386. [32] A.V. Elgersma, R.L. Zsom, W. Norde, J. Lyklema, J. Colloid Interface Sci. 152 (1992) 410. [33] R. Shastri, R.J. Roe, Org. Coating Plastic Chem. 40 (1979) 820. [34] T. Mizutani, J. Colloid Interface Sci. 82 (1981) 162.
413
[35] P. van Dulm, W. Norde, J. Lyklema, J. Colloid Interface Sci. 82 (1982) 77. [36] C. Tanford, S.A. Swanson, W.S. Shore, J. Am. Chem. Soc. 77 (1955) 6414. [37] C. Tanford, J.B. Buzzell, J. Phys. Chem. 60 (1956) 225.
Adsorption of immunogamma globulin onto various synthetic calcium hydroxyapatite particles Kazuhiko Kandori,∗ Kaori Miyagawa, and Tatsuo Ishikawa School of Chemistry, Osaka University of Education, Asahigaoka 4-698-1, Kashiwara, Osaka 582-8582, Japan Received 25 April 2003; accepted 29 January 2004
Abstract This paper presents data on adsorption of immunogamma globulin (IgG) onto synthetic rodlike calcium hydroxyapatite particles (CaHaps) with various particle lengths and calcium/phosphate (Ca/P) atomic ratios ranging from 1.54 to 1.65 and compares the obtained results to those of acidic (bovine serum albumin, BSA), neutral (myoglobin, MGB), and basic (lysozyme, LSZ) proteins reported before. The effect of electrolyte concentration on IgG adsorption was also examined. The initial rate of IgG adsorption was similar to that of BSA and was slower than that of MGB and LSZ. This fact was interpreted by the difference in the structural stability and molecular weight of these proteins. The isotherms of IgG adsorption onto the CaHap particles were of pseudo-Langmuir type. The saturated amount of adsorbed IgG IgG values (ns ) for the particles with mean particle length less than 70 nm decreased with increasing Ca/P ratio. The adsorption behavior IgG of IgG molecules was very similar to that of basic LSZ, though IgG has zero net charge. The ns value was increased with increased mean particle length of CaHaps; the relationship was less significant than that for BSA but similar to those for MGB and LSZ. The similar adsorption behavior of IgG and LSZ suggested that the Fab parts of IgG molecules preferentially adsorb onto CaHap to provide the reversed Y-shaped conformation of IgG. The change of the adsorption mode of IgG molecules from the reversed Y-shaped conformation IgG to side-on by “spreading” the Fc part of IgG molecules onto the particle surface over a longer adsorption time was suggested. The ns value was increased with increasing electrolyte concentration by screening the intra- and intermolecular electrostatic interactions of proteins. 2004 Elsevier Inc. All rights reserved. Keywords: Protein adsorption; Immunogamma globulin; Synthetic calcium hydroxyapatite particles; Bovine serum albumin; Myoglobin; Lysozyme; Electrophoretic mobility
1. Introduction Adsorption of IgG onto solid surfaces is of great interest for biomedical applications, such as immunoassays and biosensors. IgG or antibodies are proteins synthesized by an animal in response to the presence of a foreign substance called an antigen [1–3]. These proteins, naturally designed with extraordinary specificity and binding affinity for a given antigen, are used as diagnostic reagents. Therefore, many investigations have been done using different adsorbents such as silica gel [4,5], polymer latex [6–8], and Teflon [9,10] particles. In the development of diagnostic reagents, the main goal is to achieve stable attachment of IgG molecules at the solid surfaces without disrupting or masking the biological * Corresponding author.
E-mail address: [email protected] (K. Kandori). 0021-9797/$ – see front matter 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.jcis.2004.01.069
functions of the protein, such as selectivity and specificity. An IgG molecule is composed of four polypeptide chains, two heavy and two light chains, and they are grouped into two domains [1,2,11]. The molecule is composed of two Fab segments and one Fc segment and forms a “Y-shaped” conformation. Because the antigen-binding sites are located on the far ends of the Fab segments, it is desired that the IgG molecules be adsorbed onto the adsorbent surface by anchoring Fc parts with exposing the Fab segments toward to the outside. On the other hand, it is well known that calcium hydroxyapatite [Ca10 (PO4 )6 (OH)2 , CaHap] is a major inorganic component of bone and tooth and possesses a high affinity for proteins. Since CaHap is in the space group P 63 /m and its unit cell dimensions are a = b = 0.943 nm and c = 0.688 nm, possessing two different binding sites [12,13], called C and P sites on the crystal surface, it has a multiple-
K. Kandori et al. / Journal of Colloid and Interface Science 273 (2004) 406–413
site binding character for proteins. After dispersed CaHap particles in aqueous media, calcium atoms (C sites) are exposed on the CaHap surface by dissolution of OH− ions at the particle surface. Therefore, the C sites are arranged on ac or bc crystal faces in a rectangular manner with the interdistance in the a or b directions equal to 0.943 nm and the interdistance in the c direction equal to 0.344 nm (c/2). The P sites are arranged hexagonally on the ab crystal face with a minimal distance of 0.943 nm. The C sites are rich in calcium ions or positive charges and thus bind to acidic groups of proteins, but the P sites lack calcium ions or positive charges and therefore attach to basic groups of proteins. The fundamental studies of the adsorption of acidic (BSA: isoelectric point (iep) = 4.7), neutral (MGB: iep = 7.0), and basic (LSZ: iep = 11.1) proteins onto various kinds of synthetic hydroxyapatites (Hap), such as calcium Hap (CaHap), strontium Hap (SrHap), calcium–strontium Hap (CaSrHap), and carbonate-containing CaHap (Cap), have been carried out by the authors [14–20]. We clarified in these studies that the saturated amount of adsorbed BSA, nBSA , depends on the molar ratio of s cations (Ca, Sr, and/or Ca + Sr) to phosphate (cation/P) in the materials used; the nBSA increased with increase in the s cation/P ratio and was explained by the electrostatic attractive forces between negatively charged BSA and the less negatively charged CaHaps with high cation/P ratios at pH 6. The same result was further observed for the adsorption of BSA onto CaHap [14], CaSrHap [16], and Cap [17] particles. In contrast, however, the saturated amounts of adsorbed LSZ (nLSZ s ) on SrHap and CaSrHap particles decreased with increased cation/P ratio and approach zero around at cation/P = 1.70 [19], whereas various Haps exhibited constant values of nMGB and no remarkable relationship with s cation/P of the particles was detected for the neutral protein MGB [20]. The adsorption of IgG onto CaHap has been reported so far by two dental research groups. Rölla et al. [21] reported that immunoglobulins, IgA and IgG, constitute a part of the pellicle on the tooth surface. Furthermore, Sasaki et al. [22] and Eggen and Rölla [23] disclosed that the effect of NaF surface treatment of CaHap on the adsorption of IgG. To our knowledge, however, no quantitative data have been reported in the literature on the adsorption of IgG. This paper presents adsorption data of IgG onto synthetic CaHaps with various sizes and Ca/P ratios ranging from 1.54 to 1.65 and compares the results obtained to those of acidic, neutral, and basic proteins reported before. Comparison studies of these systems may contribute to further understanding of the mechanism of IgG adsorption onto CaHaps. The effect of electrolyte concentration on the IgG adsorption was also examined. The results also might be beneficial for applying CaHap to the HPLC, immunoassay, and biosensor systems.
407
Fig. 1. Electron micrographs of typical CaHap particles (Nos. 3, 6, 10 and 13) given in Table 1.
2. Experimental 2.1. Materials and methods Thirteen kinds of rodlike CaHap particles with different particle lengths ranging from 24 to 83 nm but similar diameter of 5–16 nm, as shown in Table 1, used in this study were synthesized by aging the precipitates at 100 ◦ C formed from the reaction of a solution of Ca(OH)2 with H2 PO4 in a Teflon vessel for 48 h under CO2 -free conditions. The particle length was controlled by changing the molar ratio of Ca(OH)2 to H3 PO4 in the solutions. The recipe for these particle preparations was reported in detail elsewhere [14,15]. These particles were characterized by conventional methods of XRD, TG-DTA, ICP-AES, TEM, N2 adsorption, and zeta potential measurement [14,15]. The mean particle size was obtained by measuring the length and diameter of 200 particles in each TEM micrograph. The adsorption isotherm of N2 was measured at liquid nitrogen temperatures using a computerized automatic volumetric apparatus built in house. Prior to the measurement, the samples were evacuated at 300 ◦ C for 2 h. The specific surface area was determined within a high accuracy of ±2% by fitting the adsorption isotherms to the BET equation. The typical TEM pictures of several samples are displayed in Fig. 1. It should be mentioned here again that the CaHap particles precipitated at higher concentrations of H3 PO4 (Nos. 10–13) are large rods with mean particle lengths over 70 nm. The polyclonal IgG (99% purity), fraction II and III, was purchased from Sigma Co. (G-5009). The electrophoretic mobility (EM) of the particles was measured using an electrophoresis apparatus as reported before.
408
K. Kandori et al. / Journal of Colloid and Interface Science 273 (2004) 406–413
Table 1 IgG Properties of various kinds of CaHap and saturated amounts of adsorbed IgG (ns ) IgG
No.
[H3 PO4 ] (mol)
Ca:PO4 atomic ratio
Specific surface area (m2 g−1 )
Crystallite size of (002) plane (nm)
Mean particle size (nm)
ns (mg m−2 )
1 2 3 4 5 6 7 8 9 10 11 12 13
0.24 0.24 0.24 0.24 0.24 0.235 0.24 0.24 0.247 0.249 0.252 0.265 0.25
1.54 1.56 1.58 1.59 1.60 1.62 1.62 1.64 1.58 1.58 1.59 1.61 1.65
96 91 90 101 92 94 104 104 90 97 97 112 101
46 46 40 41 23 34 38 16 29 27 17 23 24
46 × 7 36 × 8 53 × 9 69 × 8 35 × 16 33 × 7 24 × 12 29 × 13 60 × 8 83 × 7 76 × 5 83 × 7 77 × 5
0.50 0.43 0.46 0.28 0.27 0.35 0.18 0.16 0.40 0.50 0.60 0.61 0.52
2.2. Protein adsorption experiment The protein adsorption experiment was carried out in 10 cm3 Nalgen polypropylene (PP) centrifugation tubes to which varying concentrations of IgG in 1 × 10−4 mol dm−3 KCl solution of pH 6 were added to disperse the CaHaps (150 mg) by the batch method given elsewhere [14–17]. Preliminary experiments revealed that the solubility of the CaHap particles employed in the present study in KCl solution is quite low under this neutral pH condition. The calcium and phosphate ions dissolved from CaHap particles in a KCl solution, assayed by the ICP-AES method, were ca. 2.5 × 10−4 and 1.7 × 10−5 mol dm−3 , respectively. These ion concentrations are considered to be far too low for varying the solubility of IgG molecules and the electric double layer of CaHap particles. The 10 mg cm−3 IgG solutions were prepared by dissolving IgG in 1 × 10−4 mol dm−3 KCl solution. However, since IgG did not completely dissolve, the clear supernatant after the solution was centrifuged was used for the experiment. The IgG concentration in the supernatant was spectroscopically determined using an extinction of 1.45 cm3 mg−1 at 280 nm [1,4]. The initial concentrations of IgG in PP tubes were varied from 0 to 7.5 mg cm−3 . The PP centrifugation tubes were gently rotated end over end at 15 ◦ C for 48 h in a thermostat. To determine the adsorption isotherms of IgG in the static condition, the concentrations of IgG in the supernatant after centrifugation were determined by the microburet method using an UV adsorption band at 310 nm. The pHs of IgG solution before and after adsorption were measured by Toa pH-meter. In contrast, to compare the initial rates of IgG adsorption with other proteins examined before, the amounts of adsorbed IgG were measured at various incubation times from 2 to 48 h using CaHap No. 4 particles. Preliminary experiments revealed high reproducibility of the method (±2%) using the PP centrifugation tube and no protein adsorption onto the PP tube could be detected. The effect of electrolyte on the adsorption behavior of IgG was also examined by varying the KCl concentration between 1.0 × 10−4 and 1.0 mol dm−3 .
Fig. 2. Kinetics of adsorption of IgG (2), MGB (E), BSA (!), and LSZ (P) from each single-protein system onto CaHap (No. 4).
3. Results and discussion 3.1. Initial rate of IgG adsorption onto CaHap (kinetics study) The initial rate of IgG adsorption onto CaHap was examined at an initial concentration of 7.0 mg cm−3 using CaHap No. 4 as displayed in Fig. 2. For the comparison, the data of BSA, LSZ, and MGB adsorption onto the same material from each single protein system as reported in our previous studies [20] are also shown in Fig. 2 (open symbols). It is obvious that the initial rates of IgG (2) and BSA (!) adsorption are lower than those of MGB (E) and LSZ (P). The kinetic data can usually be analyzed by a diffusion-control adsorption model under quiescent conditions; protein molecules reach the particle surface by Brownian motion. Under such conditions, the rate of arrival of the protein at the particle surface is given by the Ward and Tordai equation [25], J = Cp (D/πt)1/2 , where J is the flux of the protein per unit area of particle surface, Cp is the concentration of protein, t is the time, D is the diffusion coefficient of IgG in the solution (4.0 ×10−11 m2 s−1 ), and π is 3.14. Many experimental data
K. Kandori et al. / Journal of Colloid and Interface Science 273 (2004) 406–413
confirmed diffusion-controlled adsorption from nonflowing solutions at low concentration [26,27]. However, since the solutions dispersing CaHap in PP centrifugation tubes are gently rotated end over end in a thermostat and the concentration of IgG (7.0 mg cm−3 ) is high in the present study, the diffusion-control model can not be applied to our data. On the contrary, it is well known that the proteins having low structural stability, called “soft proteins” such as IgG, BSA, and MGB, tend to change their structure upon adsorption. This structural rearrangement of IgG and BSA with high molecular weight (IgG: 146,000 Da, BSA: 67,200 Da) upon adsorption, therefore, may be the reason for the slow initial adsorption rate as compared to LSZ (14,000 Da), with high structural stability, and MGB, with low structural stability but with low molecular weight (17,800 Da). This structural change of adsorbed IgG molecules will be discussed later. Another finding in Fig. 2 is that the saturated amount of adIgG ) sorbed IgG (ns ) is much less than those of BSA (nBSA s MGB and MGB (ns ). This result indicates that the affinity of IgG to the CaHap surface is low. As we have reported before, BSA molecules are strongly adsorbed onto the CaHap, mainly through a specific electrostatic attractive force between negatively charged carboxylic groups on BSA and positively charged C sites localized on CaHap surfaces, even though the surface net charge of CaHap is negative at pH 6 [14,15]. Therefore, this specific . On electrostatic adsorption process provides a high nBSA s the other hand, since MGB molecules are expected to have a slight positive net charge at pH 6, similar specific electrostatic interaction did not operate between positively charged amino groups of MGB and C sites. The authors therefore conclude in the study that neutral MGB molecules approach the negatively charged CaHap surface much more readily than BSA ones without such specific electrostatic adsorption. In addition, since it is also well known that MGB molecules are characterized as “soft” proteins with lower structure stability, similar to BSA, we presumed that MGB could be adsorbed onto both hydrophobic and hydrophilic surfaces under attractive and repulsive electrostatic conditions, similarly to BSA. Actually, Kondo and Mihara emphasized the large conformational change of MGB upon adsorption on the colloidal particles [24]. The same mechanism can be anticipated for IgG molecules, because the IgG molecules are expected to have a slight positive net charge at pH 6, the same as MGB, as will be shown in Figs. 3 and 4 in the next section. The detailed mechanism of IgG adsorption onto CaHap surface will hereinafter be discussed. 3.2. Adsorption of IgG onto various kinds of CaHaps (static study) Figs. 3 and 4 display the characteristic adsorption isotherms of IgG and their mobilities and the pH values of the IgG solution before and after adsorption for relatively small and large CaHap particles, respectively. To avoid confusing the data points, the representative data of EM and pH for Ca-
409
Fig. 3. Isotherms of IgG adsorption onto CaHap (Nos. 1–9) (a), EM of No. 2 (b), and solution pH of No. 5 (c).
Hap Nos. 2, 5, and 13 are shown in the figures. Actually, similar data were obtained for the other CaHap particles. Here, to simplify the figure, the pH values of IgG solution before adsorption (!) are plotted at the equilibrium concentration of the corresponding solution after adsorption ("). The isotherms of IgG adsorption onto these particles are of the pseudo-Langmuir type, as we have already reported for BSA, MGB, and LSZ; the adsorption isotherms steeply rose in the lower concentration region and attained saturation [15–21]. It can be seen in Figs. 3 and 4 that the pH values of the IgG solution without CaHap particles are around 5.0–6.2, but they are increased to ca. 6.5–7.0 after adsorption onto the CaHap particle. This pH rise is due to generation of OH− ions from the CaHap particle surface by dissolution, because CaHap particles are stable under the present conditions, i.e., only produce very low contents of calcium and phosphate ions, as described before. No phosphate buffer solution was used in this study. This result therefore strongly supports the mechanism of C sites formation on the ac or bc
410
K. Kandori et al. / Journal of Colloid and Interface Science 273 (2004) 406–413
IgG
Fig. 5. Theoretical ns
values for different adsorption configurations.
Fig. 4. Isotherms of IgG adsorption onto CaHap (Nos. 10–13) (a), EM of No. 13 (b), and solution pH of No. 13 (c).
crystal surfaces was described in Section 1. The EM value shows nearly zero over the IgG concentration investigated, but the sign of the EM reverses from negative to positive with the IgG adsorption. This result indicates that the iep of IgG lies between pH 6 and 7. This fact is in agreement with the iep values of IgG reported (5.8–6.9) in other literatures [1,28,29]. The extremely small EM of CaHap particles suggests that IgG molecules can be regarded as neutral proteins under the conditions of the present study. IgG The plateau amounts of adsorbed IgG (ns ), corresponding to the saturated amounts of adsorbed IgG, are listed in Table 1. The IgG molecule has a Y-shaped conformation with dimensions of 7.0 × 6.3 × 3.1 nm3 for the Fc fragment and 8.25 × 5.0 × 3.8 nm3 for each of the two Fab fragments [2,9]. From this molecular size, we can evaluIgG ate the theoretical ns values for different adsorption configurations, i.e., end-on and side-on, by assuming that IgG molecules are adsorbed closely onto a flat surface (Fig. 5). IgG The experimental ns values obtained in the present study IgG (0.1–0.6 mg m−2 ) are less than all the theoretical ns val-
IgG
Fig. 6. (a) Plots of ns as a function of Ca/P ratio and (b) plots of IgG ns , nBSA , nMGB , and nLSZ as a function of particle length of CaHaps s s s (" > 70 nm, ! < 70 nm).
ues evaluated. Therefore, this result does not directly provide insight about the adsorption mode of IgG on CaHap. To discuss the adsorption of IgG onto CaHap in further IgG detail, we plotted the ns as a function of the Ca/P ratios of CaHap particles because our previous studies revealed that the adsorption amounts of proteins, especially for acidic and basic proteins, are strongly dependent on the Ca/P raIgG tio [19]. Fig. 6a depicts the change in the ns values with the Ca/P ratio. The data for CaHap particles with relatively large mean particle lengths over 70 nm are plotted by filled
K. Kandori et al. / Journal of Colloid and Interface Science 273 (2004) 406–413 IgG
circles. It is noteworthy that the ns values for the particles with mean particle lengths less than 70 nm (!) decrease with increased Ca/P ratio, though those for the particles over 70 nm in the mean particle length are almost constant around 0.5–0.6 mg m−2 . The authors previously found the important effect of electrostatic attractive force between protein and CaHap [18]. Since the large fraction of positively charged C sites is advantageous for the adsorption of negwas increased atively charged acidic BSA molecules, nBSA s with increased Ca/P ratios. In contrast, LSZ of basic proteins was decreased with exhibited an opposite relationship; nLSZ s increased Ca/P ratio and finally reached almost zero around Ca/P = 1.70. Unlike BSA and LSZ, no relationship between and the Ca/P value could be detected for neutral MGB. nMGB s The results observed for IgG molecules in Fig. 6a are very similar to those for basic LSZ reported before, though IgG has only slight positive charge. This unexpected but attractive result strongly suggests that the mode of IgG molecules adsorption onto CaHap is similar to that of basic LSZ ones but not to that of neutral MGB ones. IgG In contrast, the ns values for the relatively large CaHap particles (") do not follow this relationship. The authors reported the importance of C sites for the adsorption of acidic depends strongly protein BSA in the previous paper; nBSA s on the particle length, more markedly than on the Ca/P ratios, by the specific electrostatic interaction between negatively charged carboxylic acid groups of BSA and positively charged C sites on the exposed ac or bc crystal surfaces [18]. To make clear the behavior of IgG adsorption onto CaHaps, IgG therefore, we plotted the ns values as a function of the mean particle length for all the CaHap particles employed in Fig. 6b together with the results of BSA, MGB, and LSZ reported before by dotted lines [20]. It is easy to recognize IgG that the ns value is increased with increased mean particle length of CaHaps, the relationship is less significant than for BSA but similar to those for MGB and LSZ. Since IgG has a slight positive net charge as well as MGB, similar adsorption behavior can be expected on both proteins. But the slope of the line is close to LSZ, indicating that the IgG ns values are much less correlated with the C site of CaHap. This fact agrees with the results in Fig. 6a, suggesting that the neutral IgG behaves as a basic protein with high structural stability. Norde et al. reported that the adsorption behavior of the Fc domain is typically that of a “soft protein;” i.e., it adsorbs onto a hydrophilic, electrostatically repelling surface, whereas the “hard” Fab parts only adsorb if the hydrophilic silica under them is electrostatically attracted [30]. According to this result, it can be presumed that Fab parts of IgG molecules preferentially adsorb onto CaHap to take a reversed Y-shaped conformation (Fig. 5b), because IgG molecules and CaHap surfaces are electrically attracted under the conditions examined. The configuration of IgG molecules may further support this mechanism as follows: the Fab fragment of the IgG molecule is NH2 -terminal to give a positive charge, though the Fc one is COOH-terminal to provided negative charge [1,4], indicating a higher electro-
411
static attractive force between the Fab fragment and CaHap with negative net-charge. In the literature, it is inferred that IgG adsorption preferentially takes place with the Fc domain because this part is more hydrophobic and flexible than Fab part [2,8,31,32]. However, it can be assumed from the result in Fig. 6 that this mechanism cannot be applied to the adsorption of IgG onto hydrophilic and negatively charged CaHap particles at neutral pH conditions. It has also reported that adsorbed IgG molecules become more tightly bound as the contact time with the adsorbent surface increases [2,28,29]. This observation is due to slow structural rearrangements within the IgG molecules in order to optimize interactions with the adsorbent surface [1]. According to this concept, it is concluded that the conformational change is favorable for IgG, especially for the flexible Fc part. The conformational change, called “structural relaxation,” implies optimization of the protein–particle surface interactions by developing a larger number of contacts between them. Therefore, it seems reasonable to conclude that the mode of IgG molecule adsorption is changed from the reversed Y-shaped conformation to side-on within a longer adsorption time; a certain degree of “spreading” of the Fc part of the IgG molecule may take place. 3.3. Effects of electrolytes on the adsorption of IgG onto CaHap It is well known that the interactions between proteins and surface are influenced by ionic strength in the system [8,33–35]. Therefore, we examined the effects of electrolyte on the IgG adsorption onto CaHap (No. 12) by changing the concentration of KCl from 1.0×10−4 to 1.0 mol dm−3 . Prior to this experiment, we also checked the solubility of CaHap particles in these KCl solutions. However, CaHap particles exhibited high stability at all the KCl concentrations employed in this study, as described in a previous section, and no differences in calcium and phosphate concentrations and solution pH were observed under these experimental conditions. This result implies that the variation of adsorption of IgG onto CaHap is attributed to the effect of electrolytes added. The obtained adsorption isotherms measured at varied KCl concentrations are displayed in Fig. 7. There is a considerable increase in the adsorption as the KCl concentration is IgG increased up to 1.0×10−1 mol dm−3 (1), but ns is slightly −1 −3 decreased at 5.0 × 10 mol dm (filled symbols); the IgG ns exhibits a maximum at 1.0 × 10−1 mol dm−3 . Suzawa and Murakami explained the increase of nBSA onto polymer s latex with increasing electrolyte concentration by three reasons [8]: (1) The electrostatic repulsions in the interiors of protein molecules are neutralized by the addition of electrolytes, and this favors a compact protein structure. Namely, the conformational stability of protein molecules increases with increasing the ionic strength [36,37]. (2) The lateral repulsion between adsorbed protein molecules is weakened
412
K. Kandori et al. / Journal of Colloid and Interface Science 273 (2004) 406–413
Acknowledgments The authors thank Masao Fukusumi, Osaka Municipal Technical Research Institute, for help with the TEM. This work has been supported in part by a Grant-in-Aid for Scientific Research (B) from the Ministry of Education, Science, Sports and Culture.
References
Fig. 7. Isotherms of IgG adsorption onto CaHap (No. 12) at various KCl concentrations: KCI concentration: (e) 1.0 × 10−4 , (P) 1.0 × 10−3 , (!) 1.0 × 10−2 , (1) 1.0 × 10−1 , (") 5.0 × 10−1 , and (Q) 1.0 mol dm−3 .
with increase of ionic strength; thus more molecules can adsorb onto the given surface area. (3) The electrostatic repulsion between protein molecules and particle surface weakens, so protein molecules can more easily adsorb. Since IgG and CaHap are oppositely charged, reason (3) is unsuitable for explaining the present result. The appearance of maximum adsorption at 1.0 × 10−1 mol dm−3 indicates that there is an optimum KCl concentration for reasons (1) and (2). The secondary and tertiary structure of IgG could be changed at IgG higher KCl concentration and lead to depression of the ns . But the detailed mechanism of this fact is not clear at the moment and is left for subsequent papers.
4. Conclusion The initial rates of IgG and BSA adsorption are slower than those for MGB and LSZ. This fact is interpreted by the difference in the structural stability of these proteins. The IgG and BSA with less structural stability exhibit structural IgG rearrangement. The ns values for the particles with the short mean particle length are decreased with increased Ca/P ratio. This adsorption of IgG molecules is very similar to that of basic protein LSZ, though IgG has a slight positive net IgG charge as well as MGB. Therefore, the ns value is slightly increased with increased mean particle length of CaHaps, similarly to LSZ. The Fab parts of IgG molecules may preferentially adsorb onto CaHap to take the reversed Y-shaped conformation, but after a longer adsorption time the adsorption mode of IgG molecules could be changed to side-on by spreading Fc fragments of IgG molecules on the CaHap surface. The adsorption of IgG is enhanced by screening the intra- and intermolecular electrostatic interactions of protein molecules with addition of electrolyte.
[1] C.E. Giacomelli, in: A.T. Hubbard (Ed.), Encyclopedia of Surface and Colloid Science, Dekker, New York, 2002, p. 418. [2] J. Buijs, W. Norde, J.W.Th. Lichtenbelt, Langmuir 12 (1996) 1605. [3] M. Okubo, Y. Yamamoto, M. Uno, S. Kamei, T. Matsumoto, Colloid Polym. Sci. 265 (1987) 1061. [4] J.W.Th. Lichtenbelt, W.J.M. Henvelsland, M.E. Oldenzeel, R.L. Zsom, Colloids Surf. B Biointerfaces 1 (1993) 75. [5] A. Kamyshny, O. Toledano, S. Magdassi, Colloids Surf. B Biointerfaces 13 (1999) 187. [6] R. van Erp, Y.E.M. Linders, A.P.G. van Sommersen, T.C.J. Gribnau, J. Immunol. Methods 152 (1992) 191. [7] M. Okubo, Y. Yamamoto, M. Uno, S. Kamei, T. Matsumoto, Colloid Polym. Sci. 265 (1987) 1061. [8] T. Suzawa, T. Murakami, J. Colloid Interface Sci. 28 (1980) 266. [9] A.W.P. Vermeer, M.G.E.G. Bremer, W. Norde, Biochim. Biophys. Acta 1425 (1998) 1. [10] A.W.P. Vermeer, C.E. Giacomelli, W. Norde, Biochim. Biophys. Acta 1526 (2001) 61. [11] G.M. Edelman, in: J. Sterzl, I. Riha (Eds.), Development Aspects of Antibody Formation and Structure, Academic Press, New York, 1970, p. 381. [12] T. Kawasaki, M. Niikura, S. Takahashi, W. Kobayashi, Biochem. Int. 13 (1986) 969. [13] T. Kawasaki, K. Ikeda, S. Takahashi, Y. Kuboki, Eur. J. Biochem. 155 (1986) 249. [14] K. Kandori, S. Sawai, Y. Yamamoto, H. Saito, T. Ishikawa, Colloids Surf. 68 (1992) 283. [15] K. Kandori, Y. Yamamoto, H. Saito, T. Ishikawa, Colloids Surf. A 80 (1993) 287. [16] K. Kandori, M. Saito, H. Saito, A. Yasukawa, T. Ishikawa, Colloids Surf. A 94 (1995) 225. [17] K. Kandori, M. Saito, T. Takebe, A. Yasukawa, T. Ishikawa, J. Colloid Interface Sci. 174 (1995) 124. [18] K. Kandori, T. Shimizu, A. Yasukawa, T. Ishikawa, Colloids Surf. B 5 (1995) 81. [19] K. Kandori, N. Horigami, H. Kobayashi, A. Yasukawa, T. Ishikawa, J. Colloid Interface Sci. 191 (1997) 498. [20] K. Kandori, A. Fudo, T. Ishikawa, Phys. Chem. Chem. Phys. 2 (2000) 2015. [21] G. Rölla, J.E. Ciardi, W.B. Bowen, J. Dent. Res. 91 (1983) 186. [22] K. Sasaki, T. Mori, M. Terashita, M. Nagano, C. Uchiyama, Jpn. J. Oral Biol. 26 (1984) 1362. [23] H. Eggen, G. Rölla, J. Dent. Res. 91 (1983) 347. [24] A. Kondo, J. Mihara, J. Colloid Interface Sci. 177 (1996) 214. [25] A.F.D. Ward, L. Tordai, J. Chem. Phys. 14 (1946) 453. [26] G.A. Bornzin, I.F. Miller, J. Colloid Interface Sci. 86 (1982) 539. [27] P. Van Dulm, W. Norde, J. Colloid Interface Sci. 91 (1983) 248. [28] A. Martin, J. Puig, F. Galisteo, J. Serra, R. Hidalgo-Alvarez, J. Dispers. Sci. Technol. 13 (1992) 399. [29] J. Buijs, J.W.Th. Lichtenbelt, W. Norde, J. Lyklema, Colloids Surf. B Biointerfaces 5 (1995) 11. [30] W. Norde, in: J.M. Asna (Ed.), Polymeric Dispersion, Principles and Applications, Kluwer Academic, Dordrecht, 1997, p. 541.
K. Kandori et al. / Journal of Colloid and Interface Science 273 (2004) 406–413
[31] M.E. Soderquist, A.G. Walton, J. Colloid Interface Sci. 75 (1980) 386. [32] A.V. Elgersma, R.L. Zsom, W. Norde, J. Lyklema, J. Colloid Interface Sci. 152 (1992) 410. [33] R. Shastri, R.J. Roe, Org. Coating Plastic Chem. 40 (1979) 820. [34] T. Mizutani, J. Colloid Interface Sci. 82 (1981) 162.
413
[35] P. van Dulm, W. Norde, J. Lyklema, J. Colloid Interface Sci. 82 (1982) 77. [36] C. Tanford, S.A. Swanson, W.S. Shore, J. Am. Chem. Soc. 77 (1955) 6414. [37] C. Tanford, J.B. Buzzell, J. Phys. Chem. 60 (1956) 225.