Self-assembly of the sodium salts of fatty acids into limpid hydrogels through non-covalent interactions with peptides

Abstract

It is crucial for hydrogel preparation to find appropriate gelators. Fatty acids are widely distributed in nature and thus their utilization has received much attention. In this study we demonstrated that poly(α,L-lysine) with a polymerization degree of 5 (PLL₅) can bind to the sodium salts of fatty acids (SFA) with intermediate sizes, such as sodium laurate (SL), through non-covalent interactions. Such bindings resulted in SL polymerization. TEM analyses revealed that long, thin fibers initially formed upon mixing PLL₅ with SL, becoming thick and entangled after 12 h, and finally leading to the formation of limpid hydrogels. In contrast, other SFA could not form such hydrogels under the same experimental conditions. Except for PLL₅, PLL₇ also induced SL self-assembly into similar limpid hydrogels, while its other analogues with smaller or larger sizes did not. Thus, the formation of the limpid hydrogel presented here exhibited high selectivity for the size of both SFA and poly(α,L-lysine). These findings might be beneficial for the application of fatty acids in industry.

---

Introduction

Hydrogels are super-water-absorbent natural or synthetic polymer systems that can pass the “inversion test”—turn a vial of the hydrogel upside down and it is able to support its own weight without falling down.¹ Though there are several other definitions, hydrogels are essentially two-component mixtures of a small amount (typically less than 2% by weight) usually of a polymer gelator forming a network immobilizing a much larger mass of water by surface tension or capillary force.² The three-dimensional networks of hydrogels are usually formed by cross-linking polymer chains through covalent bonds, hydrogen bonding, van der Waals interactions or physical entanglements.^3,4 By imitating protein and self-assembly materials studies,^5,6 Estroff and Hamilton² broke down a hydrogel into a primary, secondary and tertiary structure for the better understanding of its formation mechanism. The primary structure of the hydrogel is determined by the molecular level recognition events that promote anisotropic aggregation in one or two dimensions of the gelator molecules. And the morphology of the aggregates is the secondary structure, namely micelles, vesicles, fibers, ribbons or sheets, which is directly influenced by the primary structure. The tertiary structure of the hydrogel is defined as the interaction of individual aggregates, which determines whether a hydrogel can be formed.

Networks immobilizing a mass of water endow hydrogels with the ability to protect drugs or functional factors from hostile environments and to release them in response to environmental stimuli such as changes of temperature or pH. Because of these interesting properties, hydrogels have been studied extensively over the past twenty years in the applications of drug or nutraceutical delivery,^7–9 tissue engineering (hydrogels are used as 3D scaffolds to support cultured cells like an extracellular matrix),¹⁰ wound healing¹¹ and contact lenses¹² (the earliest application of hydrogels). To date, different gelators have been used to fabricate hydrogel networks that contain synthetic polymers, such as poly(ethylene oxide)¹³ (PEO), and naturally occurring polymers, such as polysaccharides (including chitosan and pectin)^14,15 and proteins.¹⁶ Synthetic hydrogelators generally require a covalently cross-linking step to form such hydrogel networks. This step is usually carried out by the incorporation of a bifunctional monomer with redox initiated chemistry or UV radiation which could produce radicals that will do harm to the human body.¹⁷ In contrast, an alternative way used to form the networks is to make use of polysaccharides or proteins that can be held together (self-assembly) to form the networks by non-covalent forces without producing any radicals. Consequently, hydrogels formed by these natural components are usually thermally reversible. However, to the best of our knowledge, so far hydrogels comprised of sodium salts of fatty acids and peptides have never been reported.

Fatty acids occur in almost all living systems where they are free to participate in the biosynthesis pathway of phospholipids and of fat-reserve of cells. Despite fatty acids being widely used in the food and chemical industry, dispersing fatty acids in aqueous solutions is an interesting challenge in the search for environmentally safer surfactants from renewable sources. On the other hand, peptides represent another rich resource, also widely distributed in nature and exist in various foods. Poly(α,L-lysine) (PLL) is a good example of a water-soluble polymer with positive charges based on the naturally occurring amino acid monomer lysine, which has been reported for practical applications such as anti-bacterial agents, DNA compaction and cell transfection.¹⁸ It has been well-known that the physicochemical property of hydrogels is tightly associated with gelators.¹⁷ Therefore, it is important to find different gelators rather than the above mentioned macromolecules to prepare hydrogels for different purposes. In this study we found that PLL₅ (Fig. 1A) can bind to sodium laurate (SL) (Fig. 1B) and further induce SL polymerization to form hydrogels at an extremely low concentration. This finding represents the first case of preparation of hydrogels by sodium salts of fatty acids and peptides.

Fig. 1 Chemical structures of poly(α,L-lysine) with different polymerization degrees (3, 5, 7, 10 and 15) (A) and sodium salts of different fatty acids (B) used in this study.

Material and methods

Materials

Sodium stearate (SS), sodium palmitate (SP), sodium laurate (SL), sodium decanoate (SD) and sodium caprylate (SC) were purchased from Tokyo Chemical Industry (Tokyo, Japan). All polypeptides used in this study are linear polymers where each amino acid residue participates in two peptide bonds and is linked to its neighbours in a head-to-tail fashion rather than forming branched chains. Poly(α,L-lysines) (PLL) with different polymerization degrees, such as 3, 5, 7, 10 and 15, were purchased from Scilight Biotechnology, LLC, China. The purity of all of them is higher than 98%. Others are of reagent grade or purer. All solutions used here were filtered through a drainage membrane filter of 0.45 μm.

Isothermal titration calorimetry (ITC) analyses

Isothermal titration calorimetry (ITC) experiments were carried out at pH 7.40 and 25 °C on a Nano-ITC II Instrument (TA Instrument) as previously described.¹⁹ Prior to analysis, SL and PLL₅ were dissolved in 10 mM Tris buffer at pH 7.40, respectively. Before each titration, all solutions were degassed thoroughly under vacuum. The solution in the sample cell was stirred at 150 rpm to ensure the rapid mixing of the titrant upon injection. Titrations were performed by an automated sequence of 25 injections, each 2.0 μL PLL₅ solution was titrated into a sample cell containing 300 μL SL solution at 4.17 mM, spaced at 280 s intervals to ensure complete equilibration. A background titration, consisting of the same titrate solution but only the buffer solution in the sample cell, was subtracted from each experimental titration to account for heat of dilution. Three titrations were carried out for each measurement, and the reported experimental values are the average of individual best-fit values, fitting with independent binding model in the Nano Analyze software provided by TA Instruments. Raw data were processed using Origin 8.1 software (Originlab).

Polymerization kinetics

Polymerization of SL induced by PLL₅ was monitored as a function of time by light scattering measurements using a Cary Eclipse spectrofluorimeter (Varian) as previously described.²⁰ Both excitation and emission wavelengths were set to 450 nm, the excitation slit was set to 2.5 nm and the emission slit was also set to 2.5 nm, and the time-dependent change in scattering light was set to a 90° angle, perpendicular to the beam.

Gelation

7.50 mM PLL₅ or 5.46 mM PLL₇ in Milli-Q water was added to a series of SL solutions in water with different SL/PLL_{5 or 7} molar ratios (10 : 1, 14 : 1, 18 : 1, 20 : 1, 22.5 : 1, 25 : 1 and 50 : 1), respectively. After mixing to uniformity, the resulting mixtures were left standing at 25 °C for 12 h.

Transmission electron microscope (TEM) analyses

A drop of sample was placed on a carbon-coated copper grids for 2 min, and excess sample was removed with filter paper. The resulting samples were stained using 2% uranyl acetate for 2 min. TEM images were obtained at 30 kV through a Hitachi H-7650B transmission electron microscope.

Circular dichroism (CD) spectra analyses

SL-PLL₅ samples at different standing times were freeze-dried. The dry samples were re-dissolved to 0.05 mg mL⁻¹ in Milli-Q water before recording on a Chirascan plus spectrometer (Applied Photophysics Ltd) at room temperature (25 ± 1 °C) under a constant flow of nitrogen gas. Each sample was scanned in the range of 185–260 nm. CD spectrum was obtained as the average of three scans with the water background subtracted.

Statistical analyses

All data analysis was performed using Origin 8.1 software (OriginLab). All experiments were carried out three times.

Results and discussion

ITC measurements

Calorimeters have now been developed that are capable of accurately and reproducibly measuring the heat flow associated with the interaction of biological molecules in dilute aqueous solutions. In particular, isothermal titration calorimetry (ITC) is being used as a powerful tool to provide detailed thermodynamic information about the binding interaction between molecules. To determine whether or not poly(α,L-lysine) with a polymerization degree of 5 (PLL₅) can bind to SL, ITC experiments were carried out. The raw ITC data for a titration of sodium laurate (SL) with PLL₅ at pH 7.40 and 25 °C are shown in Fig. 2A. The integrated heats for each injection are shown after subtraction of the control injection in buffer alone. The positive peaks seen in Fig. 2A are indicative of an endothermic reaction. The heats for each injection decreased with an increase in the molar ratio of PLL₅ to SL, and there was little heat change in the end, indicating that the reaction was complete under the present conditions. The ITC data were best fitted to an independent binding model (Fig. 2B), giving an apparent equilibrium constant K = (1.15 ± 0.30) × 10⁵ M⁻¹ with the PLL₅/SL reaction stoichiometry of 0.019 ± 0.001. The apparent binding constant falls in the range of non-covalent binding, indicating that PLL₅ interacts with SL through non-covalent forces.²¹ Based on the reaction stoichiometry, it can be calculated that one PLL₅ molecule can react with almost fifty SL molecules, suggesting that polymerization reactions occur in this system.

Fig. 2 ITC measurement of SL binding to PLL₅ at pH 7.40 (10 mM Tris) and 25 °C. (A) Raw data obtained from continuous injections of 2.0 μL PLL₅ solutions at 0.75 mM into 4.17 mM SL solutions. (B) Titration plot derived from the integrated heats of binding of raw data, corrected for the heat of dilution.

As listed in Table 1, a negative value of ΔG° (−28.89 ± 0.61 kJ mol⁻¹) indicates that the polymerization reaction of SL and PLL₅ occurs spontaneously. In contrast, ΔS° (510.18 ± 5.90 J mol⁻¹ K⁻¹) is positive, indicating that the addition of PLL₅ disordered the molecules in the whole system. Thus, this polymerization reaction is driven by entropy. This result is contrary to a recent report where sodium dodecyl sulfate (SDS) was found to bind to PLL₅ with a negative ΔS° and a negative ΔH°,²² and thus the reaction appears to be driven by enthalpy. Both SDS and SL molecules have a long hydrophobic tail lauryl group, but the hydrophilic head of SDS is a sulfate group while it is a carboxyl group for SL. Such a small difference in structure between SDS and SL could result in different reaction mechanisms.

Table 1 Thermodynamic parameters for SL binding to PLL₅ or PLL₇ by ITC measurements in 10 mM Tris buffer, pH 7.40, 25 °C^a

Sample	K (M^−1)	ΔH° (kJ mol^−1)	n	ΔG° (kJ mol^−1)	ΔS °(J mol^−1 K^−1)
a The reported thermodynamic quantities are apparent values. Standards errors from replicate determinations are indicated.
PLL5	(1.15 ± 0.30) × 10^5	123.22 ± 2.01	0.019 ± 0.001	−28.89 ± 0.61	510.18 ± 5.90
PLL7	(3.97 ± 0.55) × 10^5	107.76 ± 3.86	0.013 ± 0.001	−31.96 ± 0.43	468.49 ± 7.56

---

Meanwhile, the reactions between other sodium salts of fatty acids, such as sodium stearate (SS), sodium palmitate (SP), sodium decanoate (SD) and sodium caprylate (SC), and PLL₅ were also carried out by ITC, but none of them could bind to PLL₅ within the 125 min time frame of the experiment. These results demonstrated that the reactions between PLL₅ and the sodium salts of fatty acids exhibit high selectivity, namely only the sodium of a fatty acid with an intermediate size such as SL could bind to PLL₅.

Polymerization of PLL₅ and SL

Laser light scattering is a widely applied technique for studying different molecules in solution because it can provide information about the size and conformation of the molecules and their aggregation state as well as their ability to crystallize.^23,24 To obtain more information on the reaction between PLL₅ and SL in aqueous solution, we conducted light scattering experiments at lower concentrations than the critical micelle concentration (cmc) of SL at 25 °C. As shown in Fig. 3A, the polymerization of SL was triggered by the addition of PLL₅ as indicated by the light scattering intensity rapidly increasing at 0.25 min and then reaching a plateau after about 0.30 min. It is known that surfactant molecules can polymerize to form micelles above the cmc. But in fact, micelles can be formed at different concentrations as the cmc changes according to experimental conditions. In this study, SL micelles were formed at a value of 0.95 mM in the presence of PLL₅ as indicated by light scattering experiments (Fig. 3A). The present value is much lower than the cmc of SL (∼24 mM at 25 °C),²⁵ indicating that PLL₅ has a marked effect on SL polymerization. In contrast, light scattering intensity did not increase within 1 min when mixing SS, SP, SD and SC with PLL₅ (only the result of SC is shown in Fig. 3B), indicating that, under the present experimental conditions, PLL₅ cannot induce the polymerization of the molecules with a longer or shorter hydrophobic tail. These results are in good agreement with the above observation with ITC (Fig. 2).

Fig. 3 A change in scattered light intensity of SL (A) and SC (B) in the presence of PLL₅ as a function of time at 25 °C. Conditions: both excitation wavelength and emission wavelength are 450 nm, the angle of scattering light was set at a 90° to the incident beam; [SL] = 0.94 mM, [SC] = 2.0 mM, [PLL₅] = 0.75 mM. The arrow in (B) represents the moment of PLL₅ addition.

Hydrogel formation

Subsequently we tested whether the SFA–PLL₅ mixture could gelate with increasing reaction time as previously described.²⁶ Upon adding PLL₅ to SL with different molar ratios from 10 : 1 to 50 : 1, followed by standing for 12 h, it was found that only mixtures with a SL/PLL₅ ratio of 20 : 1 to 25 : 1 formed limpid hydrogels. As shown in the left of Fig. 4A in which the ratio of SL to PLL₅ is 22.5 : 1, the SL solution alone on the left was pellucid within 12 h. In contrast, flocks appeared with the addition of PLL₅ (in the middle of Fig. 4A) to the SL solution, but they disappeared after shaking turning homogeneous. Interestingly, the resulting solutions were transformed into gels after being kept for ∼12 h at 25 °C (on the right of Fig. 4A). The formed hydrogels are very sensitive to temperature and can be completely melted at 37 °C (not shown).

Fig. 4 Photos of SL-PLL₅ hydrogels and SS–PLL₅ precipitate. (A) the left picture is a photo of SL alone, the picture in the middle is a photo immediately after the addition of PLL₅ to SL, and the right picture is a photo of limpid hydrogels during the inversion test, which was performed 12 h after the addition of PLL₅ to SL; (B) the left picture is a photo after the addition of PLL₅ to SS, and the right one corresponds to a photo 12 h after the addition of PLL₅ to SS. Conditions: [SL] = 16.78 mM, [SS] = 1.35 mM, [PLL₅] = 0.75 mM, 25 °C. All quoted concentrations are the final concentrations.

Although the hydrogels with SL as the gelator were first reported in 1941 by Marton et al.,²⁷ the concentration of SL (251.91 mM) used in the hydrogels is much higher than that in this work (16.87 mM). Moreover, the hydrogels presented herein are limpid rather than turbid as shown previously.^26,27 These results again emphasize the important role of PLL₅ in SL gelation at such a low concentration, which might be derived from the interaction between PLL₅ and SL in aqueous solution as suggested by the ITC results (Fig. 2).

In contrast, only precipitates but not limpid hydrogels were generated upon treatment of SS, SP, SD, and SC with PLL₅ under the same experimental conditions (SS–PLL₅ mixture is shown in Fig. 4B), demonstrating that the formation of the above hydrogels has high selectivity for the size of these sodium salts of fatty acids, agreeing with the above ITC and light scattering results.

TEM analyses

It is generally considered that hydrogels are composed of entangled fibers or ribbons trapping a mass of water. As expected, fibers were indeed observed with TEM in both solution and gel samples. Upon adding PLL₅ (7.5 mM) to the SL solution (16.87 mM), long, thin and flexible fibers with sharpened edges and a width of ∼35 nm (Fig. 5A) were initially generated in the mixture. After a 12 h incubation at 25 °C these fibers became very thick and intertwined with each other (Fig. 5B). The width of the fibers in the hydrogel sample (∼250 nm) was about 7 times wider than that in the solution sample (∼35 nm). It is reasonable to believe that such a large increase in the width of fibers might result in the formation of transparent hydrogels. Thus, the occurrence of long and thin fibers appears to be an early event for the formation of the SL hydrogels induced by PLL₅.

Fig. 5 TEM images of SL fibers induced by PLL₅. (A) The solution sample of SL-PLL₅. (B) The gel sample of SL-PLL₅. The inset is the enlargement in the focus box. Conditions: [SL] = 16.78 mM, [PLL₅] = 0.75 mM. All quoted concentrations are final concentrations.

As mentioned above, similar to other SFA, SS could not form hydrogels upon mixing with PLL₅, but such mixture resulted in the formation of precipitates after a 12 h incubation (Fig. 4B). A possible reason why a mixture of SS and PLL₅ cannot form hydrogels could lie in the results of the TEM analyses (Fig. S1†) showing that much shorter fibers with sharpened edges were found in the SS–PLL₅ solution, and their width is ∼250 nm, a value being ∼7 times wider than that of fibers formed in the SL and PLL₅ solution. The difference in the morphology of fibers in the SS–PLL₅ and SL-PLL₅ solutions may lead to their completely different abilities to form hydrogels.

Although the phenomenon of gelation is thought to arise from fibers or ribbons, gelation is a very complicated process. According to previous studies,^2,5,7 the hydrogel structure contains three levels: the primary, secondary and tertiary structure. Both SL-PLL₅ and SS–PLL₅ can form fibers which corresponds to the secondary structure of the hydrogel. Whether or not the secondary structure can be transformed into the tertiary structure depends on the interaction between fibers. It is evident that fibers that are long, thin and flexible are better than short, thick analogues in trapping solvent, finally facilitating the formation of gelation.² Comparing the fibers of SL-PLL₅ (Fig. 5A) with those of SS–PLL₅ (Fig. S1†) it can be found that the fibers formed by SS–PLL₅ are much shorter and thicker, and this might be an important reason why the SS–PLL₅ system can produce only precipitates rather than hydrogels under the same experimental conditions. In contrast, the long and thin SL-PLL₅ fibers could contribute a more superficial area to trap water by capillary force and surface tension. Further support for this idea was provided by a recent study of the mixing of SDS with PLL₅ that likewise resulted in the formation of shorter and thicker fibers, finally producing precipitates but not hydrogels.²²

A change in the secondary structure of PLL₅ during hydrogel formation

To shed light on the role of PLL₅ in the formation of hydrogels, we investigated a change in the secondary structure of PLL₅ upon its mixing with SL for 0, 2 and 12 h, with PLL₅ alone as a control sample, and the results are shown in Fig. 6. As expected, PLL₅ alone in water has one large negative ellipticity in the far-UV spectrum at approximately 196 nm (curve a), indicative of a typical random coil conformation.²² This observation is in good agreement with the structure of PLL₅ where the side chains of PLL₅ are all positively charged and repel each other so that the molecule is linear.¹⁸ In contrast, upon mixing PLL₅ with SL (curve b) a significantly different CD spectrum was observed, which is characterized by a much smaller negative peak at ∼198 nm, suggestive of a large decrease in random coil secondary structure of PLL₅ molecules. Interestingly, such a decrease became larger and larger with increasing reaction time from 2 h (curve c) to 12 h (curve d), suggesting that PLL₅ molecules become more ordered after interacting with SL. This might be an important reason why PLL₅ can induce SL polymerization into hydrogels.

Fig. 6 CD spectra of PLL₅ upon reacting with SL for different times. [PLL₅] = 0.05 mg mL⁻¹. The molar ratio of SL/PLL₅ was 22.5 : 1.

Effect of length size of poly(α,L-lysines) on the hydrogels

The above observation raises an interesting question of whether other PLL molecules with different length sizes would induce SL to form such hydrogels. To answer this question, PLL molecules with less or higher polymerization degrees, such as 3 (PLL₃), 7 (PLL₇), 10 (PLL₁₀) and 15 (PLL₁₅), were likewise incubated with SL under the same experimental conditions as PLL₅. It was found that only PLL₇ can react with SL as suggested by ITC shown in Fig. S2,† and a similar binding constant (3.97 ± 0.55) × 10⁵ M⁻¹ and other parameters were obtained (Table 1). In contrast, no reaction was observed between each of the other three analogues (PLL₃, PLL₁₀, and PLL₁₅) and SL (data not shown) under the same conditions. As expected, just like PLL₅ plus SL, a limpid hydrogel was also produced with the addition of PLL₇ to SL (Fig. S3†), while no hydrogel was formed upon adding PLL₃, PLL₁₀ and PLL₁₅ to SL. The same results were obtained when changing the PLL₃ or PLL₁₀/SL ratio (data not shown). Therefore, the polymerization reaction between PLL and SL likewise has a strict requirement for the size of PLL; namely, only the PLL with an intermediate size such as PLL₅ or PLL₇ is able to facilitate SL self-assembly into hydrogels.

Based on these results we proposed a four-step pathway corresponding to the formation of hydrogels by PLL₅ or PLL₇ and SL. The first step is a binding reaction of PLL₅ or PLL₇ to SL by non-covalent interactions (Fig. 2). This binding reaction is followed by a rearrangement process in which the secondary structure of PLL molecules becomes more ordered (Fig. 6). This represents step 2. Such an arrangement could result in the generation of thin, long nanofibers (Fig. 5), corresponding to step 3. After a long maturation process (∼12 h), these nanofibers became thick and intertwined with each other, leading to the final gelation (Fig. 4). This corresponds to the last step. A detailed mechanism is under investigation.

Conclusions

The present work demonstrates that sodium salts of fatty acids (SFA) can bind with poly(α,L-lysine) through non-covalent interactions at a fast rate, but such binding exhibits high selectivity for SFA; namely, only SFA with an intermediate size hydrophobic tail are able to bind to poly(α,L-lysine) also having an intermediate length size. More importantly, the binding leads to the self-assembly of sodium laurate at an extremely low concentration (16.78 mM) gradually into nanofibers, and finally leads to limpid hydrogels after a ∼12 h incubation. This finding represents the first report for the formation of limpid hydrogels by sodium salts of fatty acids and peptides, which gives a clue to making use of fatty acids, a rich resource in nature.

Abbreviations

cmc	Critical micelle concentration.
ITC	Isothermal titration calorimetry.
PLL5	NH2-(α,L-lysine)5-COOH.
PLL7	NH2-(α,L-lysine)7-COOH.
PLL3	NH2-(α,L-lysine)3-COOH.
PLL10	NH2-(α,L-lysine)10-COOH.
PLL15	NH2-(α,L-lysine)15-COOH.
SC	Sodium caprylate.
SD	Sodium decanoate.
SFA	Sodium fatty acids.
SL	Sodium laurate.
SS	Sodium stearate.

Acknowledgements

The authors gratefully acknowledge the financial support of the China High-Tech (863) Project (2013AA102208-4).

References

1. M. O. M. Piepenbrock, G. O. Lloyd, N. Clarke and J. W. Steed, Chem. Rev., 2009, 110, 1960.
2. L. A. Estroff and A. D. Hamilton, Chem. Rev., 2004, 104, 1201.
3. S. B. Ross-Murphy, Polym. Gels Networks, 1994, 2, 229.
4. K. R. Kamath and K. Park, Adv. Drug Delivery Rev., 1993, 11, 59.
5. A. Aggeli, I. A. Nyrkova, M. Bell, R. Harding, L. Carrick, T. C. McLeish, A. N. Semenov and N. Boden, Proc. Natl. Acad. Sci. U. S. A., 2001, 98, 11857.
6. B. A. Simmons, C. E. Taylor, F. A. Landis, V. T. John, G. L. McPherson, D. K. Schwartz and R. Moore, J. Am. Chem. Soc., 2001, 123, 2414.
7. J. Naskar, G. Palui and A. Banerjee, J. Phys. Chem. B, 2009, 113, 11787.
8. Y. Nagai, L. D. Unsworth, S. Koutsopoulos and S. Zhang, J. Controlled Release, 2006, 115, 18.
9. Y. Qiu and K. Park, Adv. Drug Delivery Rev., 2012, 64, 49.
10. E. S. Place, J. H. George, C. K. Williams and M. M. Stevens, Chem. Soc. Rev., 2009, 38, 1139.
11. F. H. Epstein, A. J. Singer and R. A. F. Clark, N. Engl. J. Med., 1999, 341, 738.
12. O. Wichterle and D. Lim, Nature, 1960, 185, 117.
13. Y. C. Nho, Y. M. Lim and Y. M. Lee, Radiat. Phys. Chem., 2004, 71, 239.
14. J. Berger, M. Reist, J. M. Mayer, O. Felt and R. Gurny, Eur. J. Pharm. Biopharm., 2004, 57, 35.
15. H. R. Moreira, F. Munarin, R. Gentilini, L. Visai, P. L. Granja, M. C. Tanzi and P. Petrini, Carbohydr. Polym., 2014, 103, 339.
16. L. Chen, G. E. Remondetto and M. Subirade, Trends Food Sci. Technol., 2006, 17, 272.
17. D. J. Adams and P. D. Topham, Soft Matter, 2010, 6, 3707.
18. P. Symonds, J. C. Murray, A. C. Hunter, G. Debska, A. Szewczyk and S. M. Moghimi, FEBS Lett., 2005, 579, 6191.
19. S. Yang, G. Bai, L. Chen, Q. Shen, X. Diao and G. Zhao, Food Chem., 2014, 157, 302.
20. P. A. Janmey, J. Peetermans, K. S. Zaner, T. P. Stossel and T. J. Tanaka, Biol. Chem., 1986, 261, 8357.
21. M. Hoernke, C. Schwieger, A. Kerth and A. Blume, Biochim. Biophys. Acta, Biomembr., 2012, 1818, 1663.
22. S. Zhang, S. Yang, J. Zang, R. Yang, G. Zhao and C. Xu, Chem. Commun., 2014, 50, 9943.
23. C. Li, X. Fu, X. Qi, X. Hu, N. D. Chasteen and G. Zhao, J. Biol. Chem., 2009, 284, 16743.
24. G. P. Pal, J. S. Elce and Z. Jia, J. Biol. Chem., 2001, 276, 47233.
25. A. Campbell and G. Lakshminarayanan, Can. J. Chem., 1965, 43, 1729.
26. Z. Yuan, W. Lu, W. Liu and J. Hao, Soft Matter, 2008, 4, 1639.
27. L. Marton, J. McBain and R. Vold, J. Am. Chem. Soc., 1941, 63, 1990.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra09625a

---