Isothermal Titration Calorimetry Directly Measures the Selective Swelling of Block Copolymer Vesicles in the Presence of Organic Acid.

Block copolymer (BCP) vesicles loaded with drug molecules may have a nonidentical swelling behavior due to the strong interactions between BCP vesicles and loaded molecules. A thermodynamic study of the swelling for such a system is of great importance in clarifying their pH-gated drug delivery behavior. In this study, the selective swelling of polystyrene-block-poly(2-vinylpyridine) (PS-b-P2VP) vesicles in the presence of different acids was compared using dynamic light scattering, zeta-potential, and isothermal titration calorimetry (ITC) measurements. Transmission electron microscopy observation verified that these PS-b-P2VP vesicles were mainly multilamellar. Importantly, using the ITC measurement, we first compared the thermodynamic parameters, including ΔH, ΔG, and ΔS, association binding sites (N), and binding association constants (K a) in the selective swelling of the PS-b-P2VP vesicles in low pH (pH ∼3.5), with or without a hydrogen bonding interaction. We observed that the existence of a hydrogen bonding interaction between tartaric acid/malic acid and PS-b-P2VP generates a limitation to the selective swelling of PS-b-P2VP vesicles, in which conditions will depend on the molecular structures of the organic acids and PS-b-P2VP. This work first provides a quantitative insight on the swelling of BCP vesicles in the presence of hydrogen bonding and highlights the power of ITC measurements for investigating the structural transformation of polymer nanostructures.

Introduction

Self-assembled nanostructures of block copolymers (BCPs) are widely applied in catalysis,[1,2] pharmaceutical sciences,[3] and other fields.[4−6] The interaction between BCP nanostructures and small molecules is the critical issue that needs to be addressed in these applications. For example, in a BCP-based drug delivery system, the interaction between BCP and drug molecules directly relates to their loading efficiency and the drug release triggered by the H+-gated swelling of BCP microspheres.[7] In some cases, if a drug molecule connects with polymer through hydrogen bonding or other strong interactions, the existence of such a strong interaction will make the pH-triggered swelling of polymer microspheres nonidentical; that is, the swelling will be influenced by two intertwined powers: the chain’s expansion and the contraction because of the complexation. Many characterizations, such as potentiometric titration,[8,9] dynamic light scattering (DLS),[8,9] electrophoretic mobility,[8] turbidimetric titration,[10] and electron microscopy,[10,11] have been utilized to identify the swelling of polymer microparticles. However, a complex case such as the above-mentioned nonidentical swelling is difficult to clarify via a simple characterization. Because the thermodynamic parameters as the binding sites, the binding association constants, and the detailed enthalpic and entropic contributions are of great importance for interpreting the complex swelling of polymer microspheres, a precise and direct measurement of these factors is required.

Isothermal titration calorimetry (ITC) has been extensively used in the study of biological and synthetic systems, including cell biology,[12] food chemistry,[13] drug delivery,[14] enzyme kinetics,[15] and protein adsorption,[16,17] either alone or as a supplement to other technologies. In a typical ITC experiment, the heat change is accurately measured when a solution of one compound (so-called ligand) is titrated into the solution of another compound in an isothermal “measurement cell”. As a result, the full thermodynamic profile (ΔG, ΔH, ΔS, Ka, and stoichiometry) for each binding interaction in an aqueous solution was determined. This measurement requires no additional functionalization of chromogenic, fluorogenic, or radioisotope-labeled ligands, showing its advantage as an ideal tool for the study of particle–molecule interactions.[18−23] However, such a powerful technique has only been utilized in a few cases of polymer micelles or colloid chemistry.[16,24−27] It has not been used in the study of the swelling of polymer nanostructures yet.

Given that ITC is a direct and concise pathway for studying the interaction between colloids and small molecules, in this study, we demonstrated its ability to interpret the complex swelling process, in which hydrogen bonding is mixed with selective swelling. We quantitatively characterized the swelling behaviors of the diblock copolymer (PS-b-P2VP) vesicles composed of three different molecular weights in the presence of d-tartaric acid (d-TA), malic acid (MA), or HCl aqueous solution in a thermodynamic way. Based on the size change by DLS, variation of zeta-potential, and ΔG, ΔH, and ΔS values obtained from ITC measurement, we observed that the hydrogen bonding between the pyridine N in P2VP and the carboxyl group of tartaric acid or malic acid decreases both the size expansion and surface charge of BCP vesicles as well as the enthalpy value in the swelling process, compared with that induced by HCl solution at the same pH condition. The restriction degree varies depending on the molecular structures of organic acid and the BCP. In addition, the presence of organic acids induces a distinct difference in the entropy change, binding sites, and binding association constants, compared with those in HCl-induced swelling. This study shows that the ITC measurement can provide quantitative data of the structural change of the BCP self-assembled system, which will greatly benefit the mechanism study of the swelling of BCP vesicles under the influence of various interactions.

Materials and Methods

Materials

Diblock copolymers polystyrene-block-poly(2-vinylpyridine) with different molecular weights, PS47k-b-P2VP24k (Mw/Mn = 1.07), PS30k-b-P2VP8.5k (Mw/Mn = 1.06), and PS48.5k-b-P2VP14.5k (Mw/Mn = 1.07) were purchased from Polymer Source, Inc., Canada. These subscripts for every PS-b-P2VP sample represent the molecular weights for each block. d-TA (purity >99%) and MA (purity >99%) were purchased from Sigma-Aldrich. HCl (36–38 wt % concentration) was purchased from Beijing Chemical Works. Acetone (purity ≥99.5%) was purchased from Modern Oriental Technology Development Co., Ltd. (Beijing). All chemicals were used as received without further purification.

Preparation of the Ligand Solutions and BCP Suspensions

d-TA or MA was dissolved in deionized water (Millipore Q; >18 MΩ·cm) to prepare the d-TA/MA solution (0.01 M, pH 2.5). The concentration of the aqueous HCl solution was 3.16 × 10–3 M, with a pH of 2.50. Diblock copolymers were directly dissolved in acetone (1.0 mg·mL–1) and stirred for at least 5 days at room temperature. Based on the molecular weights of three BCP samples, the molar concentrations of 2-vinylpyridine (2VP) were 2.10 × 10–4 M (PS30k-b-P2VP8.5k), 2.19 × 10–4 M (PS48.5k-b-P2VP14.5k), and 3.22 × 10–4 M (PS47k-b-P2VP24k). In the formation of BCP vesicles, 1 mL of deionized water was dropped into 1 mL of PS-b-P2VP/acetone solution at a rate of 50 μL·min–1 using a syringe pump under magnetic stirring at 150 rpm. The solution was in a closed container to avoid the evaporation of solvents in the mixing process. Then the mixed solution was dispersed in 10 mL of water to freeze the BCP nanostructures, and the obtained suspensions were kept at ambient temperature over 3 days to evaporate acetone thoroughly. After the evaporation, the concentration of BCP in the above suspension was adjusted to 0.1 mg/mL, and it was used in ITC characterization. Transmission electron microscopy (TEM) characterization demonstrated that the obtained BCP nanostructures had mainly multilamellar vesicle morphology.

Characterizations

A transmission electron microscope (TEM, H-7650B) was operated at an acceleration voltage of 80 kV. The dilute samples were deposited on copper grids, washed with deionized water three times, and dried in a vacuum oven at 30 °C overnight before TEM characterization. To determine the result of the swelling size change and zeta-potential change of the BCP nanoparticles, different amounts of HCl/d-TA/MA solution were added into 1 mL of BCP nanoparticle suspensions for 1 h, and then the suspensions were diluted and characterized with DLS (Malvern Nano-ZS90 ZetaSizer) at 25 °C.

ITC measurements were performed at 25 °C (298 K) with MicroCal PEAQ-ITC (Malvern Instruments Ltd.) in the following two steps. In the first step, the thermodynamics in the dilution of the aqueous ligand solution was measured to identify the heat change. In brief, an aqueous solution of HCl (pH ∼2.5) or d-TA (0.01 M, pH ∼2.5) and MA (0.01 M, pH ∼2.5) in the syringe was added into the sample cell of deionized water (200 μL volume). Then a similar titration was conducted in the sample cell filled with a dispersion of PS-b-P2VP vesicles (c = 0.1 mg/mL, 200 μL). In a typical titration, 0.1 μL of the titrant (HCl/d-TA/MA) is first added into the solution (200 μL) in the cell, and then 15 μL of the titrant was titrated 30 times in intervals of 200 s under stirring at 750 rpm. A reference cell was filled with deionized water. For the accuracy of the ITC measurements, each system process was repeated at least three times under the same conditions, and the data were in good agreement under the fixed concentration of BCP and acids. The curves were fitted using MicroCal PEAQ-ITC analysis software provided by Malvern Instruments to determine the enthalpy changes, binding constants, and binding ratio. The “one set of identical sites” model[19,28] has been used to calculate the stoichiometry, ΔH, −TΔS, and ΔG. Details of the ITC theory are found in the Supporting Information.

Results and Discussion

The morphologies of the BCP nanostructures used in the study were vesicles of different sizes, as presented in Figure a–c. The size distributions of these vesicles before titration were measured by DLS (Figure d). The average sizes are 412 ± 5 nm (PS30k-b-P2VP8.5k), 310 ± 12 nm (PS48.5k-b-P2VP14.5k), and 232 ± 3 nm (PS47k-b-P2VP24k). The size of PS-b-P2VP vesicles decreases with the increased molecular weight of the P2VP block. It is similar to the observation reported by Perevyazko et al.,[29] in which they demonstrated that the molecular weight of hydrophilic block has a significant influence on the size of the final nanostructures. The hydrophilic block with higher molecular weight will form a larger repulsion force among each other because of the electrostatic interaction, thus hindering the aggregation of block copolymers.

Figure 1

(a–c) TEM images of the self-assembled vesicles for (a) PS30k-b-P2VP8.5k, (b) PS48.5k-b-P2VP14.5k, and (c) PS47k-b-P2VP24k. The scale bars are 100 nm. (d) Average size of three vesicles measured by DLS. (e) Illustration of the possible two effects in selective swelling of PS-b-P2VP vesicles with or without a hydrogen bonding interaction. If HCl solution was titrated, the protonation of PS-b-P2VP happens (a), whereas if d-TA/MA was used, both the protonation (a) and the complexation (b) of PS-b-P2VP coexist.

Because of the pyridine groups in P2VP, the different degrees of selective swelling or the stretching of P2VP chains happen in the presence of acids, induced by the total effect of the protonation, hydrogen bonding, or ionic interaction. Basically, in the HCl-induced stretching of P2VP chains, the protonation is the dominant factor; however, for the d-TA- or MA-induced swelling, hydrogen bonding interactions between P2VP and TA (or MA) will be involved (Figure e).[30]

Adding acids into PS-b-P2VP vesicles will induce the change of their size and surface charge, which is directly related to the pH condition of swollen particles. In these ITC titrations, the concentration was 3.16 × 10–3 M for HCl and 0.01 M for d-TA/MA solution. The pH values of these three acids were all 2.5. Based on the preliminary tests, we observed that the titration of ITC was conducted by adding 15.1 μL of HCl/d-TA/MA into 200 μL of BCP suspension (0.1 mg/mL). We measured the pH values of these titrated systems from the start to the final state and found that they are highly in accord with each other (Figure a). The pH values in the final mixtures were the same, ∼3.50. Meanwhile, the zeta-potential of these vesicles with addition of acids was also measured (Figure a). The zeta-potential for three BCP particles increased by adding HCl, showing the neat protonation effect. However, in the MA titration cases, the zeta-potential slightly increased at the beginning of the titration and then decreased, whereas in the d-TA series, the surface charge decreased continually, showing a strong influence by the hydrogen bonding interaction (Figure a). The hydrodynamic diameters of the above three BCP samples increased after adding HCl, d-TA, or MA, which exactly verified the stretching of P2VP chain-induced size expansion of the vesicles in all cases (Figure b). If the swelling is only dominated by the pH value, the addition of different acids will induce a similar size expansion because they were in a similar pH condition (∼3.5). However, the size change of PS-b-P2VP vesicles in the presence of HCl, tartaric acid, or malic acid presents a distinct difference (Figure b). The volume increase (ΔV/V) of PS47K-b-P2VP24K vesicles induced by HCl was 40.9%, which was significantly larger than those induced by d-TA/MA (28–29%). This tendency was maintained in all three vesicles (Figure b). It indicated that, in addition to the protonation effect that merely enlarges the vesicle’s size like in the HCl case, a counteraction which can limit the size’s increase emerged with the addition of d-TA or MA, which is due to their hydrogen bonding interaction (Figure e). These size and zeta-potential changes with addition of acids demonstrates that, except for the pH condition, the type of acids added also impacts the selective swelling behavior of BCP vesicles.

Figure 2

(a) Changes of pH (histogram) and zeta-potential (line chart) after addition of different amounts of HCl/d-TA/MA in three PS-b-P2VP dispersions. (b) Average size of three vesicles after adding 15.1 μL of acid for 1 h was measured by DLS. The size changes of vesicles (ΔV/V) in different systems are presented. The concentration is 3.16 × 10–3 M for HCl and 0.01 M for d-TA/MA solution. The volume of BCP dispersions was 200 μL, with a concentration of 0.1 mg/mL.

To study the selective swelling thermodynamics of PS-b-P2VP vesicles with or without hydrogen bonding, ITC has been employed to measure the heat flow, whereas HCl/d-TA/MA aqueous solution was gradually dropped into the vesicle’s suspensions. In these titrations, the heat flow in ITC mainly originated from two contributions: the first is the dilution heat of the acid being diluted in water, and the other is due to the protonation/complexation of vesicles. In the complexation, the ionic interactions and hydrogen bonding interactions are mixed. We measured the dilution heat of the three acid solutions into the water by ITC (Figure S1), and the results were used as a reference for the following measurements in acid/BCP titrations to perform the curve-fitting analysis. Figure a,b and Figure S2 show the heat flow curves together with the corresponding isotherm of PS47k-b-P2VP24k vesicles with HCl, d-TA, and MA. These curves for PS30k-b-P2VP8.5k and PS48.5k-b-P2VP14.5k are presented in Figure a,b and Figure S3. In the study of BCP swelling by ITC, the interpretation of the heat flow data is a big challenge. We have made some simplifications in analyzing these data. First, the total amount of P2VP was used in the calculation of the molar ratio in these isotherms, although only a part of P2VP in the BCP nanoparticles was protonated with the addition of acid. The detailed calculation is shown in the discussion part of the Supporting Materials. Thus, the molar ratio on the x-axis of all isotherms indicates the molar ratio of additional acids to that of P2VP in BCP vesicles in the measurement cell. Second, in the absence of stoichiometry in the BCP swelling, identifying an appropriate model to interpret titration curves is a great challenge.[31] There are several different models demonstrated in the literature,[32,33] including one set of site models,[19,34] multiple noninteracting binding site (N) models,[18,35] Langmuir isotherms,[33] and small molecule nanoparticle models.[36] As we have mentioned, for HCl, d-TA, and MA cases, except for dilution heat, the common factor contributing to the heat flow in titration was the protonation/complexation of BCP, in which different complexation degrees were involved. Thus, we tentatively proposed a simplified model for the acid-induced selective swelling of BCP in the following equations:where K is the reaction constant, [2VP] is the concentration of 2VP in the unchanged BCP vesicles in the suspension, [HA] is the acid concentration, and [HA–2VP] is the swollen 2VP in the BCP, which represents the protonation/complexation of BCP vesicles. In the case of HCl, [HA–2VP] is the protonated 2VP in the BCP, whereas in d-TA or MA titrations, the 2VP–HA complex may become dominant.

Figure 3

ITC profiles for the titration of the (a) HCl solution and (b) MA solution into a dispersion of PS47k-b-P2VP24k vesicles in water at 298 K. The x-axis in the integrated heat data presents the molar ratio of adding acid versus that of the 2VP block (3.22 × 10–4 M) in PS-b-P2VP vesicles in the measurement cell. (c) Comparison of the free energy change (ΔG), enthalpy change (ΔH), and entropy change (expressed as −TΔS) in the titration of PS47k-b-P2VP24k vesicles with three acids.

Figure 4

ITC profiles for the titration of the MA solution into the dispersion of (a) PS30k-b-P2VP8.5k and (b) PS48.5k-b-P2VP14.5k vesicles in water at 298 K. The x-axis in the integrated heat data presents the molar ratio of adding acid versus that of the 2VP block in BCP nanoparticles in the measurement cell. The concentrations of 2VP were 2.10 × 10–4 M (PS30k-b-P2VP8.5k) and 2.19 × 10–4 M (PS48.5k-b-P2VP14.5k). (c,d) Comparison of the free energy change (ΔG), enthalpy change (ΔH), and entropy change (expressed as -TΔS) in the titration with three acids, (c) PS30k-b-P2VP8.5k and (d) PS48.5k-b-P2VP14.5k, respectively.

d-TA and MA are weak organic acids, and the ionization of these two acids is different from that of HCl. However, we omitted the difference in these acids. Here, we simply consider the swelling of BCP nanoparticles as the neat result of protonation and complexation of BCP induced by adding acids, whereas in the HCl case, the protonation is dominant; in d-TA or MA cases, the complexation based on hydrogen bonding or ionic interactions offsets the protonation somehow. Based on this hypothesis, we chose the one set of site model to interpret these data. In the literature of thermochemical studies about cation exchange in CdSe nanocrystals via ITC, Rioux et al. also chose the one set of site model to fit their curves.[19,28] The gradual cation exchange from CdSe to AgSe nanocrystals may share some similarities with the protonation of PS-b-P2VP nanoparticles with the addition of acids.

Based on the independent binding curve fit analysis, the enthalpy changes (ΔH) were obtained (Figure c and Figure c,d). ΔH in all cases was negative, indicating that the selective swelling of P2VP in the presence of different acids is exothermic. We noticed that the type of BCPs (i.e., the volume fraction of the hydrophilic P2VP chains) influences the enthalpy changes in the swelling. The ΔH values for PS47k-b-P2VP24k, no matter in the titration by HCl, d-TA, or MA, are more negative than the other two BCPs, which may be because it has P2VP chains that are longer than those in the other two samples. When the d-TA or MA solution was used as the titrant, the exothermic values were lower than that in the HCl titration for three BCP cases (Figure c), showing this enthalpy change is related to the complexation degree of acids. Given that the pKa values of d-TA and MA are 3.07[36] and 3.40,[37] respectively, and d-TA has one more carboxyl group than MA, d-TA has a relatively stronger tendency to form a H bond with P2VP, corresponding to the relatively lower enthalpy change in the titration.

In addition, it is noticed that the entropy change of these titration processes in HCl titration is negative. Rubinstein et al. indicated that the stretching of a coiled chain of polymers will subsequently lose its entropy, which is caused by the reduced number of possible conformations.[38] In addition, we thought here that the entropy decrease was because a large number of H3O+ in solution was tied to the BCP vesicles in the protonation, inducing the entropy decrease of the system. The increasing hydrophilicity of P2VP after protonation also contributes to the entropy decrease because of the more water molecules being bound with the swollen P2VP chains. However, in the MA or d-TA titration, the entropy showed different changes depending on the BCP composition. For PS47k-b-P2VP24k, the titration maintained the entropy decrease, with the order of HCl > MA > d-TA in the absolute value of ΔS. The smaller value of ΔS indicates the weakened protonation and hydrophilicity of P2VP in the case of weak acids. The zeta-potential value of PS47k-b-P2VP24k vesicles decreased with the addition of d-TA (Figure a), confirming the decrease of their protonation. In the literature, the complex formation based on a hydrogen bonding interaction indeed weakens the ability to bind water.[37] Meanwhile, for hydrated MA or d-TA, these water molecules bound with d-TA or MA in their aqueous solutions were released in the formation of complexes, leading to an entropy increase. Therefore, an entropy decrease induced by protonation and hydrophilicity is offset by the entropy increase from water release in the d-TA or MA titration. For PS30k-b-P2VP8.5k and PS48.5k-b-P2VP14.5k with lower volume fractions of P2VP, the protonation was lower, and the entropy increase induced by the releasing water in P2VP/d-TA complexing became dominant, giving a net increase in entropy. In addition, the ionic interactions between protonated P2VP and acid ions may also have a contribution to the entropy change. However, its portion is hardly identified in the current study.

Moreover, all three systems show the negative Gibbs free energy in the selective swelling, declaring a spontaneous reaction for all systems. Comparing the contributions from ΔH and −TΔS, we can consider that the selective swelling process was mainly driven by the enthalpy change. We have further compared the binding sites (N) and the binding association constants (Ka) for all PS-b-P2VP systems (Table ) to identify the relationship between the thermodynamic parameters and the type of acids. As we have indicated above, the dilution of acids was subtracted before the calculation for these thermodynamic parameters. The N as the binding sites of acids with polymers presents the sum of the protonation and complexation. However, in this condition (pH ∼3.5 in all systems after titration), the 2VP of vesicles was protonated insufficiently for the pKa of P2VP to be ∼3.[39] Thus, the contribution from protonation was limited. For the d-TA/MA cases, these d-TA/MA molecules can form complexes with 2VP via hydrogen bonding, leaving the larger N sites in MA/TA cases. Moreover, the glassy PS of multilamellar vesicles restricted the access of acid into the interior of the vesicles. As a result, only part of the 2VP groups was occupied at the binding sites. Thus, the value of N should be less than 1, especially for the HCl cases. The binding sites for all three BCP series increased from HCl to MA to d-TA, which is following the complexation degree via hydrogen bonding. In HCl systems, the N is much less than that in d-TA/MA systems because of the lack of hydrogen bonding. Moreover, the binding association constants Ka shows a tendency of PS30k-b-P2VP8.5k > PS48.5k-b-P2VP14.5k > PS47k-b-P2VP24k, which coincides with the order of their average size. The large size of vesicles may create a superior probability for organic acid molecules to bind. In this way, because it is hard to maintain the same particles’ size in different ITC measurements, the thermodynamic parameters obtained from different ITC measurements showed a slight change for the same BCP/acid system, but this fluctuation cannot change the tendency related to the type of acids and BCP composition. Generally speaking, this work paves a way for extending the ITC measurement in the study of complex swelling of various micelles, such as liposomes or thermoresponsive block copolymer micelles. The thermodynamics parameters obtained by the ITC method will deepen our understanding of the swelling behaviors of these complex systems.

Table 1

Binding Association Constants (Ka) and Binding Sites (N) for Three Vesicles Composed of PS-b-P2VP with Different Molecular Weights in the Titration of HCl, d-TA, or MA Solution

		PS30k-b-P2VP8.5k	PS48.5k-b-P2VP14.5k	PS47k-b-P2VP24k
Ka (×105 M–1)	HCl	1.59 ± 0.66	1.04 ± 0.14	1.03 ± 0.43
	MA	2.33 ± 0.38	1.89 ± 0.70	0.57 ± 0.14
	d-TA	1.35 ± 0.51	0.43 ± 0.12	0.07 ± 0.01
binding sites N	HCl	0.12 ± 0.01	0.08 ± 0.01	0.03 ± 0.01
	MA	0.52 ± 0.01	0.45 ± 0.01	0.32 ± 0.03
	d-TA	0.58 ± 0.02	0.58 ± 0.04	0.39 ± 0.06

Conclusions

In this study, we have yielded new insights into the selective swelling thermodynamics of PS-b-P2VP vesicles in the presence of organic acids (tartaric acid and malic acid) through ITC measurement. Based on the ITC data, we found that the selective swelling of PS-b-P2VP was an enthalpically driven process, no matter with or without the hydrogen bonding interactions. The lower exothermic values for d-TA/MA titrations compared to that in the HCl titration for three BCP samples showed that the enthalpy change was determined by the protonation, which was offset by the complexation based on hydrogen bonding interaction or ionic interaction. In addition, the protonation and accompanying hydrophilicity of P2VP increased the bonding water on polymer chains, leading to the entropy decrease; this effect was weakened in the d-TA or MA titration due to the less protonation and was even exceeded by the entropy increase from water release in the d-TA/MA-BCP complexation. Furthermore, these thermodynamic parameters, such as binding site (N) and binding association constants (Ka), obtained from the ITC measurement presented interesting details of the relationship between selective swelling and the molecular structure, and morphological features of BCP vesicles, such as Ka, demonstrated a tendency of PS30k-b-P2VP8.5k> PS48.5k-b-P2VP14.5k > PS47k-b-P2VP24k that coincides with the order of the average size of the vesicles. Combined with the knowledge about the size and zeta-potential change of BCP vesicles in selective swelling, these characteristically thermodynamic profiles based on ITC measurements shed light on the structural rearrangement of polymer particles in selective swelling. These new observations enrich our understanding of the selective swelling of BCP vesicles under the influence of a strong interaction, which will benefit the development of a delicately controlled drug delivery system based on the pH-triggered size change of BCP particles.