Multiple Steps and Critical Behaviors of the ... - ACS Publications

Feb 17, 2007 - Tooba Shoaib , Ariel Carmichael , R. E. Corman , Yun Shen , Thanh H. ... Linda Pravinata , Mahmood Akhtar , Phillip J. Bentley , Thanis...
2 downloads 0 Views 382KB Size
Download PDF
2456

J. Phys. Chem. B 2007, 111, 2456-2462

Multiple Steps and Critical Behaviors of the Binding of Calcium to Alginate Yapeng Fang,*,† Saphwan Al-Assaf,† Glyn O. Phillips,†,‡ Katsuyoshi Nishinari,†,§ Takahiro Funami,| Peter A. Williams,⊥ and Liangbin Li# Glyn O. Phillips Hydrocolloid Research Center, North East Wales Institute, Plas Coch, Mold Road, Wrexham LL11 2AW, UK, Phillips Hydrocolloid Research Ltd., 45 Old Bond Street, London W1S 4AQ, UK, Department of Food and Human Health Sciences, Graduate School of Human Life Science, Osaka City UniVersity, Sugimoto, Sumiyoshi, Osaka 558-8585, Japan, Hydrocolloid Laboratory, San-Ei Gen F.F.I., Inc., 1-1-11 Sanwa-cho, Toyonaka, Osaka 561-8588, Japan, Centre for Water-Soluble Polymers, North East Wales Institute, Plas Coch, Mold Road, Wrexham LL11 2AW, UK, and National Synchrotron Radiation Laboratory and Department of Polymer Science and Engineering, UniVersity of Science and Technology of China, Anhui, 230026, China ReceiVed: December 28, 2006; In Final Form: January 19, 2007

Previous research on the binding and gelation of calcium/alginate in aqueous solution were mostly conducted in the (semi-)concentrated regime, and it did not provide details of the binding process and the formation of egg-box junctions. In the present investigation, the binding of calcium to alginate, of low and high molecular weight and different guluronate/mannuronate ratios, was investigated in dilute solutions using isothermal titration calorimetry (ITC), Ca2+-selective potentiometry, and viscometry techniques. The results reveal three distinct and successive steps in the binding of calcium to alginate with increased concentration of Ca ions. They were assigned to (i) interaction of Ca2+ with a single guluronate unit forming monocomplexes; (ii) propagation and formation of egg-box dimers via pairing of these monocomplexes; and (iii) lateral association of the egg-box dimers, generating multimers. The third step has different association modes depending on the molecular weight of alginate. The boundaries between these steps are reasonably critical, and they closely correlate with the Ca/guluronate stoichiometry expected for egg-box dimers and multimers with 2/1 helical chains. The formation of egg-box dimers and their subsequent association are thermodynamically equivalent processes and can be fitted by a model of independent binding sites. The binding of Ca to alginates of different guluronate contents is controlled by a balance between enthalpy and entropy.

Introduction Alginate is a collective term for a family of exopolysaccharides produced from brown seaweeds and some bacteria. Chemically they are linear copolymers of (1f4)-linked β-Dmannuronate (M) and R-L-guluronate (G) residues. The residues are arranged in a blockwise pattern with G blocks and M blocks interspersed with MG alternating blocks.1 The overall composition of M/G residues and their distribution patterns vary with seaweed species. However, very recently, the availability of C-5 epimerases enabled the tuning of the fine structure of alginate and tailoring the composition and distribution patterns of M/G blocks.2 Alginate has been widely used in food, biomedical, pharmaceutical, and sewage-treating industries.1,3-5 Most of the applications are relevant to its ability to bind divalent cations such as alkaline earth metals and heavy metals and to form gels. It has been demonstrated that alkaline earth metals preferentially bind toward the G residues of alginate rather than M residues,6,7 and their binding affinities increase in the order Mg , Ca < Sr < Ba.8 * Author to whom correspondence should be addressed. Tel: +44 (0) 1978 29 3330. E-mail: [email protected]. † Glyn O. Phillips Hydrocolloid Research Center, North East Wales Institute. ‡ Phillips Hydrocolloid Research Ltd. § Osaka City University. | San-Ei Gen F.F.I., Inc. ⊥ Centre for Water-Soluble Polymers, North East Wales Institute. # University of Science and Technology of China.

The “egg-box model”, first proposed by Rees et al., has been generally accepted to describe the formation of alginate gels in the presence of alkaline earth metals.9-11 The model indicates that the G blocks along alginate chain adopt a 2/1 helical conformation, so forming buckled regions. Divalent cations, such as Ca2+, are coordinated within the cavities created by a pair of the buckled G sequences, forming an egg-box dimer (Figure 1, parts a and b). Stokke et al. employed a small-angle X-ray scattering technique to study the gelation of alginate during the in-situ release of Ca2+ from Ca-EGTA or CaCO3 using a slow-acidifying lactone.12-15 In addition to the principal formation of egg-box dimers at lower concentrations of Ca2+, they found a lateral association of the egg-box dimers at higher concentrations of Ca2+ (as shown in Figure 1c). Although the 2/1 helix is the most probable polymer conformation involved in egg-box junction zones, the possibility of the existence of 3/1 helical conformation cannot be completely ruled out.7,16 The Ca/G stoichiometry for egg-box dimers and multimers is directly related to this still-debated issue. According to a molecular modeling analysis,7 2/1 helical conformation yields the Ca/G stoichiometry of 1/4 and 1/2 for dimers and infinite multimers, whereas 3/1 helical conformation substantially increases the stoichiometry to 1/3 and 2/3, respectively. With regard to the details of divalent cation binding and eggbox structure formation, Siew et al. investigated the binding of Mn2+ and Ca2+ to alginate by using electron spin resonance (ESR) spectroscopy.17 To interpret the binding behaviors in the

10.1021/jp0689870 CCC: $37.00 © 2007 American Chemical Society Published on Web 02/17/2007

Multi-Step Binding of Calcium to Alginate

Figure 1. Schematic representation of the hierarchical structure of eggbox junction zones in alginate/calcium gels: (a) coordination of Ca2+ in a cavity created by a pair of guluronate sequences along alginate chains; (b) egg-box dimer, and (c) laterally associated egg-box multimer. The black solid circles represent the oxygen atoms possibly involved in the coordination with Ca2+. The open circles represent Ca2+ ions.

framework of counterion condensation theory, they suggested a monocomplexation process (i.e., charge reversal) preceding the formation of egg-box dimers.17 This proposed mechanism, however, needs to be confirmed experimentally. Donati et al. very recently devised a theoretical model to describe the binding of Ca2+ to alginate/pectin and chain associations from the aspects of both specific interaction and counterion condensation,18-20 and they applied it successfully to the experimental data obtained for pectin.20 At variance, this model assumes that the egg-box dimers grow progressively rather than via an allor-none process or having any induction period.19 Apart from these investigations, there have been few attempts to examine the details of the binding of Ca2+ to alginate and to elucidate the pathway of chain-chain association. Moreover, most of the previous studies have dealt with concentrated solutions and/or gel states, which do not allow different molecular events that occur during the binding to be distinguished and provide a detailed picture on the formation of eggbox junction zones. Here, the breakdown of the binding process was achieved by stepwise addition of small amounts of CaCl2 into dilute alginate solutions, of different molecular weights and M/G ratios, followed by the investigations using high-sensitivity isothermal titration calorimetry (ITC), Ca2+-selective potentiometry, and relative viscometry. Experimental Section Materials. Two sodium-type alginates, Protanal LFR 5/60 (batch no. 911846) and Protanal H120L (batch no. 240304), with different molecular weights and M/G ratios, were obtained from FMC BioPolymer (Norway). Their molecular characteristics are the following: Protanal LFR 5/60, Mw ) 35 kDa, G% (w/w) ) 64%, M% (w/w) ) 36%; Protanal H120L, Mw ) 404 kDa, G% (w/w) ) 46%, M% (w/w) ) 54%. The two alginate samples were designated P170G64 and P2000G46, respectively, where for example 170 represents the polymerization degree and 64 the content of G residues. CaCl2‚2H2O of analytical grade was purchased from Fisher Scientific (UK), and used as the source of Ca2+. Sample Preparations. Alginates were dispersed in 20 mM acetate buffer (pH 5) and hydrated overnight with gentle shaking to ensure complete dissolution. Then, the alginate solutions were dialyzed extensively against 20 mM acetate buffer (pH 5). The buffer collected after dialysis was used to adjust alginate concentrations to the desired values and to prepare CaCl2 solution. The protocol of sample preparations adopted here is

J. Phys. Chem. B, Vol. 111, No. 10, 2007 2457 mainly to eliminate the errors that could be caused by pH and ionic strength mismatches in isothermal titration calorimetry measurements. Unless otherwise specified, all the alginate and CaCl2 solutions used in this paper were in acetate buffer. Isothermal Titration Calorimetry (ITC). A CSC 4200 isothermal titration calorimeter (Calorimetry Sciences Corporations, USA) was employed to measure enthalpies of the binding of Ca to alginate at 25 °C. Aliquots (10-µL) of 7.5 mM CaCl2 were sequentially injected into a 1300-µL reaction cell initially containing either acetate buffer or alginate solutions (0.026% and 0.052% w/w). There were a total of 24 injections for each measurement with an interval of 1000 s between two successive injections. The stirring speed was set at 297 rpm for all the experiments. ITC results were analyzed using the software BindWorks 3.1 provided by CSC. The binding isotherm obtained by integrating injection peaks could be fitted by a model of independent binding sites (eq 1):21

Q ) V∆H ×

{

[L] +

}

1 + [M]nK - x(1 + [M]nK - [L]K)2 + 4K[L] 2K (1)

where Q is accumulative heat, and V, [L], and [M] are the volume of cell, ligand concentration (which here refers to Ca concentration), and macromolecule concentration (which here refers to residue concentration), respectively. Iterative curve fitting yielded thermodynamic parameters including the binding constant K, binding enthalpy ∆H, and stoichiometry n. Relative Viscosity Measurements. The change of relative viscosity upon stepwise addition of 50 µL of 7.5 mM CaCl2 into 6500 µL of 0.052% alginate solutions was measured by using an Ubbelohde-type capillary viscometer at 25 °C. The experiments imitated the conditions as used in ITC experiments and represent the scaling up of ITC measurements. Taking into account the dilution effect involved with titration, control measurements were also done, in which acetate buffer was added into alginate solutions. The relative viscosity was calculated as follows (eq 2):

ηr ) ts/t0

(2)

where ts is the flow time of samples, and t0 is the flow time of solvent. Ca2+-Selective Potentiometry. Free Ca2+ concentration was determined using an Orion 97-20 Ionplus Ca2+-selective electrode in conjunction with an Orion 4 Star multifunctional meter (Thermo Electro Corporation, USA) at 25 °C. Calibration curves were established using 0.1 M standardized calcium solution. Likewise, the experiments used the same titration protocol as in ITC measurements. Results Isothermal Titration Calorimetry Results. Heat flow versus time profiles resulting from the sequential injection of 7.5 mM CaCl2 into the reaction cell initially filled either with 0.26% or 0.52% (w/w) P170G64 alginate solution are shown in Figure 2a,b. A blank experiment in which 7.5 mM CaCl2 was titrated into acetate buffer was also undertaken to calibrate the heat of dilution of CaCl2 (Figure 2c). The ITC thermogram for 0.026% P170G64 exhibited a two-step binding process. The injection peak first becomes smaller with addition of CaCl2, and then it turns at the third injection to grow steeply into a maximum. After passing the maximum, the injection peak decreases

2458 J. Phys. Chem. B, Vol. 111, No. 10, 2007

Fang et al.

Figure 2. ITC thermograms recorded for injecting 7.5 mM CaCl2 into (a) 0.026% alginate P170G64, (b) 0.052% alginate P170G64, and (c) acetate buffer solutions. The corresponding binding isotherms were displayed in (d) 0.026% P170G64 (b), 0.052% P170G64 (O). The solid lines represent the curve-fitting results of the second steps of the binding isotherms using a model of independent binding sites (eq 1).

TABLE 1: Binding Parameters Obtained from the Curve Fitting of the Second Steps of the Binding Isotherms for P170G64 and P2000G46 sample

na

nGb

K (103/M)

∆H (kJ/mol)

∆Gc (kJ/mol)

∆Sd (J mol-1 K-1)

Ca/P170G64 Ca/P2000G46

0.196 ( 0.021 0.202 ( 0.011

0.306 ( 0.033 0.439 ( 0.024

10.7 ( 2.6 10.6 ( 1.4

-15.0 -11.6

-23.0 -23.0

26.8 38.3

a n was obtained on the basis of the total molar concentration of M and G residues. b nG was calculated on the basis of the molar concentration of G residues. c ∆G ) -RT ln K. d ∆S ) (∆H - ∆G)/T. Here R is the gas constant, and T is the absolute temperature.

steadily. The ITC thermogram for P170G64 at a doubled concentration of 0.052% has a similar tendency, with the turning point being shifted to the sixth injection. In contrast, the titration of CaCl2 into acetate buffer generated only a very tiny amount of heat. This means that the observed two-step phenomena can be attributed to the binding of Ca2+ to alginate rather than any solvent effect. The binding isotherms obtained by integrating the injection peaks and then subtracting the dilution heat of CaCl2 are displayed in Figure 2d for 0.026% and 0.052% P170G64. We find that the second steps of the binding isotherms are well described by the model of independent binding sites (eq 1). Iterative curve fitting gave the same set of binding parameters for the two P170G64 solutions. The results and other calculated thermodynamic parameters are given in Table 1. Similarly, the ITC measurements for another alginate sample P2000G46 also show a two-step binding with Ca2+ (Figure 3). The turning points were, respectively, at the second and fourth injections for 0.26% and 0.52% (w/w) P2000G46, which ap-

peared relatively earlier in comparison with P170G64. The second steps in their binding isotherms also fit well the model of independent binding sites (eq 1). The curve-fitting results are included in Table 1. The bindings of Ca2+ to P170G64 and P2000G46 have the nearly same binding stoichiometry n against the total concentration of M and G residues. However, when calculated in terms of G residues, stoichiometry nG is larger for P2000G46. The binding constant K is almost equal for the two alginate samples, which indicates the same value of ∆G for P170G64 and P2000G46. Considering that the two samples have different values of ∆H, it may imply an enthalpy-entropy compensation principle existing for alginates with different G contents.22 When the binding isotherms of P170G64 and P2000G46 were plotted against the ratio of added Ca to G residues (R), they show a general behavior, namely, the critical boundary between the first and the second steps located at the same value of R ) 0.25 (Figure 4).

Multi-Step Binding of Calcium to Alginate

J. Phys. Chem. B, Vol. 111, No. 10, 2007 2459

Figure 3. ITC thermograms recorded for injecting 7.5 mM CaCl2 into (a) 0.026%, and (b) 0.052% alginate P2000G46 solutions. The corresponding binding isotherms were displayed in (c) 0.026% P2000G46 (b), 0.052% P2000G46 (O). The solid lines represent the curve-fitting results of the second steps of the binding isotherms using a model of independent binding sites.

Figure 4. Assignment of different steps in the binding isotherms of P170G64 and P2000G46 according to R () Ca/G ratio). Note that Step II and Step III are indistinguishable here; the illustration is to maintain consistency with Figures 5 and 6.

Relative Viscosity Data. The changes of relative viscosity ηr of P170G64 and P2000G46 upon stepwise addition of CaCl2 are shown in Figures 5 and 6, respectively. The data for two control experiments were also included for comparison. Consistent with the ITC data, ηr for P170G64 also has a turning point at R ) 0.25. Before this point, ηr decreases with increasing R; after this point, ηr increases. Furthermore, ηr shows a sharp increase at R ) 0.55. On this basis, three steps can be identified for the interaction of P170G64 with Ca2+. In Step I, ηr of P170G64 is slightly lower than for the control sample, suggesting a reduction in the individual chain size. ηr values during

Figure 5. The change of relative viscosity ηr with R () Ca/G ratio) during the titration of 7.5 mM CaCl2 into 0.052% (w/w) P170G64. The open circles show the relative viscosity for control samples titrated with buffer solvent.

Steps II and III are much larger compared with those of the control sample. This indicates a growth of chain size during these steps. Three steps are also identified in the ηr versus R curve for P2000G46 (Figure 6). Their boundaries are also around R ) 0.25 and 0.55, respectively. The difference is that P2000G46 shows a decrease in ηr during Step III. This would indicate that the molecular size decreases after the increase in Step II. Moreover, compared with P170G64, P2000G46 showed a

2460 J. Phys. Chem. B, Vol. 111, No. 10, 2007

Fang et al. that Ca2+ binds with alginate with greater difficulty at the initial stage than at the later stage.24 Additionally, three distinct regions were found in the Scatchard plot. It should be noted that the boundaries between the three regions coincide with the threestep behaviors as revealed by ITC and viscosity measurements. Discussion

Figure 6. The change of relative viscosity ηr with R () Ca/G ratio) during the titration of 7.5 mM CaCl2 into 0.052% (w/w) P2000G46. The open circles show the relative viscosity for control samples titrated with buffer solvent.

Figure 7. (a) The plot of ν against R () Ca/G residues ratio), and (b) the corresponding Scatchard plot during stepwise addition of 7.5 mM CaCl2 into 0.052% (w/w) P170G64. [Ca2+]f is the free calcium concentration as measured by ion-selective electrode. ν is the average number of calcium ions bound to one G residue.

greater decline in ηr during Step I, so indicating a more significant reduction in the individual chain size for P2000G46. Free Ca2+ Measurements. The bound Ca concentration can be derived by subtracting the free Ca2+ concentration as measured by the Ca2+-selective electrode from the total added Ca concentration. Figure 7, parts a and b, shows the plot of ν against R, and the corresponding Scatchard plot23 for P170G64. ν is the average number of Ca ions bound to one alginate G residue. The most striking effect was seen in the Scatchard plot, which overall exhibits a markedly convex shape, which implies

Multiple Steps of the Binding. As demonstrated by ITC, viscosity and Ca2+-selective electrode measurements, the binding of Ca2+ to alginate involves three distinct steps upon increasing concentration of CaCl2. This multistep binding does not appear to have been reported previously. Previous studies were either conducted in a (semi-)concentrated solution/gel state or in the way that Ca is mixed “in one go” with alginate. Under these conditions, the different molecular events corresponding to the three steps might happen simultaneously or overlap with each other. This made it impossible to break down the binding process into the three steps. In contrast, the highest concentration of alginate used in the present study is 0.052% (w/w), which is well within the dilute region.25 The CaCl2 was added progressively into alginate solutions. The viscosity measurements indicate that during Step I individual alginate chains collapse/shrink, giving a reduced viscosity. Moreover, the binding enthalpy of Step I was estimated to be -4 kJ/mol.26 This value is close to the reported enthalpy for the monocomplex formation of the carboxylic group with Ca.27 It is, therefore, reasonable to attribute Step I to the formation of monocomplex between Ca2+ and the G units of alginate (Figure 8). Since the G units of the alginate chain have a much higher specificity for calcium binding than M units,6,7 the binding of Ca2+ to M units will not be considered in the discussion. The formation of monocomplex on one hand reduces the charge density on alginate chains, and on the other hand gives rise to local charge reversal to form some positively charged patches.17 These two effects which both contribute to the decrease (increase) of intramolecular repulsion (attraction) result in a more compact conformation leading to a reduction in viscosity. P2000G46 has a relatively lower content of G than P170G64, which suggests that P2000G46 is more flexible, since polymannuronte was reported to have a lower persistence length (11.9 nm) than polyguluronate (21 nm).7 More significantly, P2000G46 has a much higher molecular weight than P170G64. Both of these lead us to conclude that P170G64 approximates to a short rigid rod whereas P2000G46 is closer to a long flexible chain. The fact that P2000G46 exhibited a more conspicuous decline in relative viscosity in Step I than P170G64 could be due to the greater flexibility of P2000G46 chains and thus giving them greater potential to collapse/shrink. Step II involves a growth in molecular size, and it is a much stronger exothermic event compared with the Step I (Figure 4). Moreover, it becomes evident only when R > 0.25, which corresponds to the Ca/G stoichiometry (0.25) of egg-box dimers composed of 2/1 helical chains (Figure 1b). Step II can thus be assigned to the formation of egg-box dimers by the pairing of monocomplexes (Figure 8). The egg-box dimers possess a onedimensional ordered/regular structure (Figure 1b).16 This feature complies with the observation of the very strong exothermicity of Step II. The characteristics of Step III vary with the alginate samples; the short rigid P170G64 shows a continuous and even steeper growth in molecular size, whereas the long flexible P2000G46 leads to a decrease (Figures 5 and 6). Nevertheless, Step III emerges around the same value of R () 0.55) for both of the two alginate samples. This R value is comparable to the Ca/G

Multi-Step Binding of Calcium to Alginate

Figure 8. Schematic illustration of the multiple-step binding of Ca2+ to alginate: (a) short-chain alginate, and (b) long-chain alginate. The zigzag lines, smooth lines, and dots stand for G blocks, M blocks, and calcium ions, respectively. R is the feeding ratio of Ca2+ to G residue.

stoichiometry (0.5) of laterally associated egg-box multimers composed of 2/1 helical chains (Figure 1c). We assign this step to the formation of egg-box multimers via lateral association of egg-box dimers. Since the P170G64 chain is quite short and rigid, the only possible way for the egg-box dimers to laterally aggregate is via inter-cluster association (Figure 8a). This results in a further growth of molecular size. In contrast, the lateral association of egg-box dimers for P2000G46 is most likely to happen only within an individual cluster (Figure 8b), because of the greater flexibility and the smaller number of clusters present (Note that at fixed concentration, the higher-molecularweight P2000G46 would form fewer clusters in Step II, compared with P170G64.). The intra-cluster association of eggbox dimers causes a reduction in molecular size. Critical Behaviors of the Binding. In addition to the observed multiple steps, another appealing feature of the binding is that the boundaries between those steps appear to be critical, particularly the one between the first two steps (Figures 2-4). The injection heat exhibits a discontinuous change (jump) at the boundary of Steps I and II (Figure 4). Relative viscosity shows a regular linear decrease in Step I, and this only reduces until Step II (Figures 5 and 6). R ) 0.25 seems to be a critical threshold to initiate Step II, and hence the formation of egg-

J. Phys. Chem. B, Vol. 111, No. 10, 2007 2461 box dimers. This value coincides with the Ca/G stoichiometry of the egg-box dimers composed of 2/1 helical chains. This indicates that unless the ratio R matches the stoichiometry of egg-box dimers, the dimerization will not be initiated. From Figure 7a, ν is around 0.1 at R ) 0.25, indicating that roughly one out of ten G residues is bound with Ca, and 60% of the added calcium still exists in free form at this stage. The bound Ca (i.e., monocomplexes) may serve as nuclei for the propagation/growth of egg-box dimers, and the free Ca could play a role in screening electrostatic repulsion between two alginate chains so that they can approach each other closely to form a dimer. Nucleation frequently initiates phase transitions which lead to the formation of ordered structures, for example, crystallizations.28 Step I, in this sense, can also be regarded as a nucleation step prior to the formation of ordered structures of egg-box dimers. Indeed, the Scatchard plot (Figure 7b) has a convex shape, which indicates that the binding in Step I is extremely difficult. This is in accordance with the nature of a nucleation process.28 There has been lack of direct evidence on whether the formation of egg-box dimers is an all-or-none process or a progressive process. Donati et al. have hypothesized in their theoretical model that an alginate/pectin chain is constantly evolving from single chains to egg-box dimers during the binding with calcium.19 However, our experimental results suggest that a nucleation mechanism (or nearly an all-or-none process) is more appropriate to describe the formation of eggbox dimers. The results support the charge reversal mechanism that has been proposed by Siew et al., in which the formation of monocomplexes precedes the initiation of egg-box dimers.17 Step III seems to appear only when R > 0.55 (Figures 5 and 6), although its boundary is not as sharp/critical as that between the first two steps. This value approximates the Ca/G stoichiometry (0.5) of laterally associated egg-box multimers composed of 2/1 helices. Stoichiometry may also be a prerequisite for the lateral association of egg-box dimers. Although Step II and Step III can be distinguished by viscometry, they cannot by ITC (Figure 4). The injection heats change smoothly when crossing through these two steps, and on the whole they could be fitted with a model of independent binding sites. This means that the formation of dimers and their subsequent association under the mediation of Ca2+ are thermodynamically equal processes. This finding queries the current understanding that the inter-dimer association is mainly governed by electrostatic interactions.7 As has been mentioned above, the observed boundaries of the three steps (R ) 0.25 and 0.55) are consistent with the stoichiometry of egg-box structures having 2/1 helical alginate chains. This indicates that the 2/1 helix is the most probable conformation involved in the alginate-calcium egg-box junctions. Thermodynamics of the Binding. ITC experiments give an apparent stoichiometry n of 0.2 for Ca2+ binding to alginate in terms of overall M and G residues (Table 1). This value agrees well with the one reported by Siew et al. in the study of Mn2+ binding to alginate.17 They pointed out that such a stoichiometry was significantly lower than the value predicted by Manning’s counterion condensation theory (0.33), and it could not be interpreted even after modifying the Manning theory by charge reversal.17 If n is converted into nG (the binding stoichiometry against G residues only), P170G64 and P2000G46 have the binding stoichiometry of 0.31 and 0.44, respectively. They are well within the range of 0.25-05 predicted from the structure of egg-box junctions (Figure 1). It may imply that the binding of Ca2+ to alginate should be described in the context of the

2462 J. Phys. Chem. B, Vol. 111, No. 10, 2007 formation of egg-box junction zones rather than purely by Manning’s counterion condensation theory. Table 1 shows that P170G64 and P2000G46 have the same binding free energy ∆G, with the latter having a smaller binding enthalpy ∆H but a larger binding entropy ∆S. It appears that an enthalpy-entropy compensation principle operates for the binding of Ca2+ to different alginates.22 This enthalpy-entropy compensation can be explained as follows: (1) the smaller binding enthalpy of P2000G46 originates from the less perfect egg-box junctions formed for a lower-G-content sample; (2) however, it is compensated with the larger binding entropy due to a weaker restriction of Ca2+ to G residues. Implications for More Concentrated Alginate Systems. Although the multistep binding mechanism discussed above is obtained for dilute alginate solutions, it also might be applicable to the binding/gelation of more concentrated alginate systems, as long as the introduction of Ca2+ is slow and homogeneous (e.g., via in-situ release of Ca2+ from Ca-EGTA or CaCO3 using slow acidifying lactone12-15). It has been well-known that the gelation of concentrated alginate solutions upon exposure to calcium ions is rapid, strong, and irreversible, resulting in inhomogeneous gels containing fish-eye defects.29 The formation of fish-eye defects is due to the inhomogeneous mixing of calcium with alginate. The multistep binding behaviors in such a fast gelation process are most likely to be indistinguishable by the conventional techniques such as rheology, and so forth. As illustrated in Figure 8, the mode of lateral association of egg-box dimers in Step III depends on the length and rigidity of alginate chains. This leads us to expect that the gel properties of alginates of low and high molecular weights, for example, syneresis, may be different. Conclusion Multiple steps in the binding of Ca2+ to alginate have been identified using ITC, Ca2+-selective potentiometry, and viscometry, which include monocomplexation, dimerization, and lateral association of alginate chains. They occur sequentially rather than simultaneously. Whether stoichiometry is matched or not seems to dictate the initiation of dimerization and lateral association, which results in rather critical boundaries. In particular, the monocomplexation between Ca2+ and single guluronate residues serves as a nucleation step that is a prerequisite to the formation of egg-box dimers. A mechanism for the binding process has been proposed, which accommodates both short rigid and long flexible alginate chains (Figure 8). The same methodology is now being applied to the binding of Ca to pectin, and the results will be published separately. Acknowledgment. Y.F. is thankful for financial support from PHRC (U.K.) and San-Ei Gen F.F.I., Inc. (Japan), and for

Fang et al. helpful discussions with Chee Kiong Siew (Centre for WaterSoluble Polymers, NEWI, U.K.). References and Notes (1) Draget, K. I.; Smidsrod, O.; SkjakBraek, G. In Polysaccharides and Polyamides in the Food Industry. Properties, Production, and Patents.; Steinbuchel, A., Rhee, S. K., Eds.; Wiley-VCH Verlag GmbH & Co. KGaA: Weinheim, 2005; p 1. (2) Donati, I.; Holtan, S.; Morch, Y. A.; Borgogna, M.; Dentini, M.; Skjak-Braek, G. Biomacromolecules 2005, 6, 1031. (3) Soonshiong, P.; Feldman, E.; Nelson, R.; Heintz, R.; Yao, Q.; Yao, Z. W.; Zheng, T. L.; Merideth, N.; Skjakbraek, G.; Espevik, T.; Smidsrod, O.; Sandford, P. Proc. Natl. Acad. Sci. U.S.A. 1993, 90, 5843. (4) Chen, J. P.; Wang, L. Sep. Sci. Technol. 2001, 36, 3617. (5) Chen, J. P.; Hong, L. A.; Wu, S. N.; Wang, L. Langmuir 2002, 18, 9413. (6) Braccini, I.; Grasso, R. P.; Perez, S. Carbohydr. Res. 1999, 317, 119. (7) Braccini, I.; Perez, S. Biomacromolecules 2001, 2, 1089. (8) Kohn, R. Pure Appl. Chem. 1975, 42, 371. (9) Grant, G. T.; Morris, E. R.; Rees, D. A.; Smith, P. J. C.; Thom, D. FEBS Lett. 1973, 32, 195. (10) Morris, E. R.; Rees, D. A.; Thom, D.; Boyd, J. Carbohydr. Res. 1978, 66, 145. (11) Thom, D.; Grant, G. T.; Morris, E. R.; Rees, D. A. Carbohydr. Res. 1982, 100, 29. (12) Stokke, B. T.; Draget, K. I.; Yuguchi, Y.; Urakawa, H.; Kajiwara, K. Macromol. Symp. 1997, 120, 91. (13) Stokke, B. T.; Draget, K. I.; Smidsrod, O.; Yuguchi, Y.; Urakawa, H.; Kajiwara, K. Macromolecules 2000, 33, 1853. (14) Yuguchi, Y.; Urakawa, H.; Kajiwara, K.; Draget, K. I.; Stokke, B. T. J. Mol. Struct. 2000, 554, 21. (15) Draget, K. I.; Stokke, B. T.; Yuguchi, Y.; Urakawa, H.; Kajiwara, K. Biomacromolecules 2003, 4, 1661. (16) Li, L. B.; Fang, Y. P.; Vreeker, R.; Appelqvist, I.; Pelan, E. Biomacromolecules, 2007, 8, 464. (17) Siew, C. K.; Williams, P. A.; Young, N. W. G. Biomacromolecules 2005, 6, 963. (18) Donati, I.; Cesaro, A.; Paoletti, S. Biomacromolecules 2006, 7, 281. (19) Donati, I.; Benegas, J. C.; Cesaro, A.; Paoletti, S. Biomacromolecules 2006, 7, 1587. (20) Donati, I.; Benegas, J. C.; Paoletti, S. Biomacromolecules 2006, 7, 3439. (21) Cited from User’s Manual for Isothermal Titration Calorimeter Model CSC 4200; Calorimetry Science Corp., U.S.A.. (22) Sinn, C. G.; Dimova, R.; Antonietti, M. Macromolecules 2004, 37, 3444. (23) Scatchard, G. Ann. N.Y. Acad. Sci. 1949, 50, 660. (24) Imai, K. J. Biol. Chem. 1974, 249, 7607. (25) Nickerson, M. T.; Paulson, A. T. Carbohydr. Polym. 2004, 58, 15. (26) The binding enthalpy of Step I was calculated for P170G64 on the basis of the data shown in Figures 4 and 7a. The total heat of the first six points in Figure 4 for P170G64, divided by the corresponding concentration of bound Ca that was read from Figure 7a, gave a binding enthalpy of about 4 kJ/mol. (27) Christensen, T.; Gooden, D. M.; Kung, J. E.; Toone, E. J. J. Am. Chem. Soc. 2003, 125, 7357. (28) Cheng, S. Z. D.; Lotz, B. Polymer 2005, 46, 8662. (29) Skjakbraek, G.; Grasdalen, H.; Smidsrod, O. Carbohydr. Polym. 1989, 10, 31.