Interactions between Chitosan and Alginate ... - ACS Publications

May 13, 2015 - Robyn Aston, Medini Wimalaratne, Aidan Brock, Gwendolyn Lawrie, and Lisbeth Grøndahl*. School of Chemistry and Molecular Biosciences, ...
0 downloads 0 Views 4MB Size
Article pubs.acs.org/Biomac

Interactions between Chitosan and Alginate Dialdehyde Biopolymers and Their Layer-by-Layer Assemblies Robyn Aston, Medini Wimalaratne, Aidan Brock, Gwendolyn Lawrie, and Lisbeth Grøndahl* School of Chemistry and Molecular Biosciences, The University of Queensland, Brisbane QLD-4072, Australia

Downloaded via TUFTS UNIV on July 9, 2018 at 08:01:39 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

S Supporting Information *

ABSTRACT: Biopolymers are researched extensively for their applications in biomaterials science and drug delivery including structures and complexes of more than one polymer. Chemical characterization of complexes formed between chitosan (CHI) and alginate dialdehyde (ADA) biopolymers established that while electrostatic interactions dominate (as determined from X-ray photoelectron spectroscopy (XPS)) covalent cross-linking between these biopolymers also contribute to their stability (evidenced from immersion in salt solution). It was furthermore found that imine bond formation could not be directly detected by any of the techniques XPS, FTIR, 1H NMR, or fluorescence. The layer-by-layer assemblies of the biopolymers formed on silica colloids, glass slides, and alginate hydrogel beads were evaluated using XPS, as well as zeta potential measurements for the silica colloids and changes to hydration properties for the hydrogels. It was found that the degree of oxidation of ADA affected the LbL assemblies in terms of a greater degree of CHI penetration observed when using the more conformationally flexible biopolymer ADA (higher degree of oxidation).



2 and C-3.8 This increased chain flexibility renders this polymer nongelling in the presence of Ca2+ ions.9 ADA has been explored for the formation of covalently cross-linked gels incorporating either stable hydrazone bonds by reaction with adipic acid dihydrazide10,11 or labile acetal and hemiacetal bonds by reaction with ALG.12 Injectable scaffolds produced from ADA and gelatin in the presence of small amounts of borax demonstrated slower drug (primaquine) release rates when using ADA of a high degree of oxidation, and this was attributed to the ability of the drug to undergo a Schiff reaction with the aldehyde groups requiring subsequent degradation (rather than simple diffusion) processes to take place in order to release the drug.13 More recently ADA has been evaluated for biological tissue fixation and found to perform as effectively as glutaraldehyde.14 The buildup of polysaccharide-based layer-by-layer (LbL) films has attracted considerable attention in biomaterials and drug delivery application as a route to producing tailored

INTRODUCTION Alginate (ALG) is used for a range of applications due to its proven biocompatibility and gelation properties under mild conditions using predominantly calcium ions, to form crosslinking junctions by diffusion from a surrounding solution1 or from an internal source.2 Within the biomedical field, ALG has received attention in wound healing, tissue engineering, and drug delivery applications due to its structural similarity to the extracellular matrices of tissues. 3 ALG is an anionic polysaccharide derived from seaweed and comprises α-Lguluronic acid units and (1−4)-linked β-D-mannuronic acid units arranged in contiguous block sequences (MMMM, GGGG, or MGMG) dictated by their source.4 ALG does not degrade in mammalian tissue and only low molecular mass fragments are cleared from the body.5 However, the oxidation of ALG using sodium periodate generates alginate dialdehyde (ADA) which is susceptible to degradation in PBS.6 Oxidation of ALG occurs at the C-2 and C-3 positions of the uronic units. The resulting aldehyde groups in ADA exist in equilibrium with their hemiacetal form.7 ADA has reduced chain stiffness compared to ALG as a result of the increased chain flexibility introduced by cleavage of the bond between C© 2015 American Chemical Society

Received: March 23, 2015 Revised: May 11, 2015 Published: May 13, 2015 1807

DOI: 10.1021/acs.biomac.5b00383 Biomacromolecules 2015, 16, 1807−1817

Article

Biomacromolecules assemblies.15 The use of polysaccharides in such LbL assemblies has been studied primarily involving the positively charged polyelectrolyte chitosan (CHI), a biopolymer that is degradable in mammalian tissue. The choice of a negatively charged polyelectrolyte to form assemblies, however, has been diverse and have included heparin,16−18 chondroitin sulfate,16 dextran sulfate,19,20 hyaluronan (HA),16,21 and alginate (ALG).22−27 The film thickness in these LbL assemblies of CHI and either ALG or HA has been shown to increase with increasing pH and increasing ionic strength.24,28 Exponential growth of the polyelectrolyte layers has been reported for (CHI/HA) multilayers when the ionic strength was 0.3 M, while linear growth was observed at very low ionic strengths (added salt of 10−4 M) and an intermediate growth was seen for ionic strength of 0.02 M.21 Furthermore, it has been established that (CHI/HA) multilayers are highly hydrated and that the intermolecular interactions are dominated by Hbonding.16 In general, exponential growth has been associated with weak intermolecular interactions as well as with charge shielding in solutions of high ionic strength. In the current study, the biodegradable biopolymers CHI and ADA were explored in polyelectrolyte complex formation and in LbL assembly processes. The degree of oxidation of ADA can be controlled, and polymers with two different degrees of oxidation (20 and 40%, e.g., 20ADA and 40ADA) were investigated. Considering the limited number of publications investigating polyelectrolyte complexes between CHI and ADA we completed a full characterization through FTIR, X-ray photoelectron spectroscopy (XPS), confocal microscopy, and NMR to evaluate the potential imine formation between these two polyelectrolytes. LbL assemblies were investigated on silica colloids, glass slides, and Ca-ALG hydrogels using XPS as well as zeta potential measurements for the silica colloids and changes to hydration properties for the Ca-ALG hydrogels.



pellet. The difference in wet mass of the PEC before and after immersion in the salt solution was determined. A range of solutions was investigated in the LbL assembly processes. CHI (1.0 or 2.0 g L−1) was dissolved either in an acetic acid/acetate solution at pH 3.7, c = 0.10 M, or in a citric acid buffer at pH 3.0, c = 0.05 M. 20ADA and 40ADA (1.0 or 2.0 g L−1) were dissolved in water resulting in a solution of pH 6.0. To the solutions was added either calcium or magnesium chloride to a cation concentration of 0.01 M. For the LbL assemblies on freshly prepared washed silica particles, 6 mL of suspended silica particles (particle concentration of 0.011 g mL−1) was added 2.5 mL of the respective polyelectrolyte solutions (2 g L−1); the optimal amount of CHI solution was gauged in preliminary experiments from zeta potential measurements. Adsorption was allowed to proceed for 20 min with intermittent shaking after which samples were then washed 3 times with Milli-Q water to remove excess polyelectrolyte. The adsorption process was continued until the required number of layers was obtained. LbL assembly was completed on glass coverslips treated by base “piranha” solution (2.5 mL of milli-Q water, 0.5 mL of 30% hydrogen peroxide, and 0.5 mL of 28% ammonia solution; heated to 80 °C for 20 min). They were then immersed in the polyelectrolyte solutions (3 mL, 1 g L−1) for 10 min followed by washing in water before applying the next layer. Spherical calcium-cross-linked alginate hydrogel beads (Ca-ALG) were prepared by extruding 5 mL of 2% w/v ALG in a syringe fitted with an 18G needle into 50 mL of a 3% w/v CaCl2·2H2O (0.20 M) solution. The cross-linking solution was stirred at a high speed until a vortex had formed and the ALG was slowly injected at an approximate height of 5 cm above the solution. After extrusion, the speed of the magnetic stirrer was decreased and the Ca-ALG hydrogels were left to continue cross-linking for 2 h. The Ca-ALG hydrogel beads were removed from the solution and washed with milli-Q water. The LbL assembly process was carried out on these beads as described for the glass coverslips. The LbL assemblies are named according to the pH of the CHI solution, the type of ADA and the presence of divalent cation, e.g., a LbL assembly of CHI dissolved in acetic acid/acetate solution and 20ADA with both solutions containing 0.01 M CaCl2 is named (CHI[3.8]/20ADA)Ca. Characterization. Attenuated total reflectance Fourier transform infrared (ATR-FTIR) spectrometry was performed on dry samples. ATR-FTIR spectra (4 scans, 4 cm−1 resolution, wavenumber range 4000−650 cm−1) were obtained using a PerkinElmer FTIR Spectrum 2000 Spectrophotometer with a diamond/ZnSe crystal window. All spectra were recorded under ambient conditions. SPECTRUM software (PerkinElmer) was used for data collection. X-ray photoelectron spectroscopy (XPS) was performed using a Kratos Axis Ultra X-ray photoelectron spectrometer using monochromated Al Kα source (1486.6 eV) at 15 kV and 10 mA (150 W). Survey scans were performed over 1200−0 eV with 1.0 eV steps at pass energy of 160 eV. Narrow scans were performed with 0.05 steps at pass energy of 20 eV with a vacuum system giving a base pressure of ∼10−8 Torr. CasaXPS software was used to calculate the atomic concentrations. The binding energy of each sample was calibrated according to the C (1s) peak at 285.0 eV.31 In the curve fitting of the N 1s narrow scans of PEC samples the amount of amide was fixed to that determined by microanalysis (see Supporting Information). Confocal laser scanning microscopy of the starting polymers and freeze-dried ALG/CHI[6.5] and 40-ADA/CHI[6.5] was conducted on a Zeiss LSM 510 META using a 20× objective lens. Bright field and fluorescence images were obtained using excitation wavelengths of 405, 488, 543, and 633 nm. Zeta-potential (ζ-potential) and particle diameter of the silica particles were determined using a Malvern Instruments Zetasizer 3000HSa. Particles were suspended in 5 mM MES buffer (I = 10 mM, pH = 7.0). ζ-Potential measurements were taken in a flow cell, and particle diameter was determined by dynamic light scattering (DLS) by placing 2.5 mL of the above suspension in a plastic cuvette in the instrument.

EXPERIMENTAL SECTION

Materials. Medium viscosity alginic acid sodium salt (ALG), derived from brown algae, was purchased from Sigma-Aldrich. Alginate dialdehyde (ADA) was prepared by established procedures12 with a degree of oxidation of 20 or 40% (e.g., 20ADA and 40ADA). Chitosan hydrochloride (CHI) was purchased from Novamatrix Ultrapure Polymer Systems. Silica particles were produced by the Stöber process, based on a method described by Bogush et al.29 Calcium chloride dihydrate (ACS grade), magnesium chloride hexahydrate (99%), sodium citrate tribasic dihydrate (99.9%), and 2-(N-morpholino)ethanesulfonic acid (MES) were sourced from Sigma-Aldrich. Citric acid anhydrous (99.5%) was from Row Scientific. Glacial acetic acid (99%) was from Univar. Sodium acetate trihydrate (99%) was from Merck. Milli-Q water was used throughout. Methods. Polyelectrolyte complexes (PECs) were prepared following the method of Wen et al.30 Briefly, 1 mg mL−1 solutions of the ALG, 20ADA, 40ADA and CHI were prepared by dissolving each polymer separately in acetic acid solution (10 mM, pH 3.8) or in water (pH 6.5) overnight at room temperature. These ALG or ADA and CHI solutions were then combined simultaneously by dropwise addition, with continuous stirring at 100−200 rpm, to an equivalent volume of acetic acid solution or water. The resulting suspension was stirred for at least 30 min before being centrifuged at 4400 rpm for 20 min to produce the PEC sediment as a white solid. The PECs were washed and dried in a desiccator. The PECs are named according to the polymer identity as well as the pH at which they are formed, e.g., 40ADA/CHI[3.8] is the PEC formed between 40ADA and CHI at a pH of 3.8. Samples of ALG/CHI[6.5] and 40ADA/CHI[6.5] were immersed in a saturated solution of sodium chloride for a period of 1 week with continuous magnetic stirring before centrifuging to reform the PEC 1808

DOI: 10.1021/acs.biomac.5b00383 Biomacromolecules 2015, 16, 1807−1817

Article

Biomacromolecules Ca-ALG beads were dried at 46 °C in air prior to nitric acid digestion in accordance with the EPA method 3051 and analysis of calcium content by ICP-OES spectrometry on a VARIAN VIST instrument. All samples were run in duplicate. Hydrated mass and bead diameter of the calcium alginate beads after each adsorbed layer was determined by measuring the wet mass of individual beads on a precision balance to an accuracy of 0.01 mg and the diameter recorded using Vernier calipers, respectively. Measurements were determined using four to eight replicates. The hydrated mass values were used to calculate water uptake (Uw) according to eq 1, while the diameter was used to calculate swelling degree (Sw) using eq 2 based on previous work.32 Uw =

Wf − Wi × 100% Wi

Table 1. Chemical Characterization by XPS and Microanalysis of CHI and the Insoluble Matrix Formed When Mixing ALG, 20ADA, or 40ADA with CHI sample

pH

NH3+/Ntotalb (±0.05)

C/Nc

CHI ALG/CHI[3.8]a 20ADA/CHI[3.8]a 40ADA/CHI[3.8]a ALG/CHI[6.5] 20ADA/CHI[6.5] 40ADA/CHI[6.5]

3.8 3.8 3.8 6.5 6.5 6.5

0.31 0.51 0.44 0.47 0.48 0.45 0.43

5.3 11.8 12.1 11.9 11.1 11.0 11.2

a

[Acetic acid] = 10 mM. bDetermined from the XPS N 1s narrow scan. cDetermined from microanalysis.

(1)

where wf is the final mass and wi the initial mass before the LbL assembly process. ⎛⎛ ⎞3 ⎞ d Sw = ⎜⎜⎜ f ⎟ − 1⎟⎟ × 100% ⎝⎝ d i ⎠ ⎠

(2)

where df is the final diameter and di the initial diameter before the LbL assembly process.



RESULTS AND DISCUSSION Characterization of Polyelectrolyte Complexes. The starting polymers CHI, ALG, 20ADA, and 40ADA were characterized by microanalysis, hydroxylamine hydrochloride titration, FTIR, and XPS to evaluate the chemical composition and chemical functionalities present. The data and further methods information are presented in the Supporting Information (Tables S1 and S2). The degree of deacetylation of CHI, determined by microanalysis, was found to be 91% (Table S1, Supporting Information). From the XPS data (Table S1, Supporting Information) a Cl/NH3+ ratio of 1.5 was observed (theoretical value Cl/NH3+ = 1). The significantly higher experimentally determined Cl/NH3+ ratio highlights an apparent underestimation of total N and/or NH3+ by XPS. Such discrepancies in the Cl/N ratio have been reported previously for hydrochloric acid salts.33,34 The degree of oxidation for the two ADA polymers were determined to be in good agreement with the theoretical values of 20 and 40% (Table S2). The O/Na ratio obtained from the XPS spectrum of ALG (Table S2, Supporting Information) correlates with the theoretical value of 6. However, a largely underestimated C− O/O−C−O/COO− ratio in ALG (and the ADA polymers, Table S2, Supporting Information) of 16:7:1 compared to the theoretical 4:1:1 prevents the use of the narrow scan COO−/ NH3+ ratio to accurately evaluate the extent of electrostatic interactions in the PECs. Because of the underestimation in components of both the C 1s and N 1s peaks in the ALG and CHI polysaccharides, the NH3+/Ntotal ratio is used to judge electrostatic interaction in the PECs (described below). Microanalysis of the PECs yielded a C/N ratio of approximately 12 for all samples produced (Table 1). This correlates with one ALG (or ADA) monomer per CHI monomer. Incorporation of acetate ions from the acetic acid solution in place of ALG or ADA would have resulted in a lower C/N value. In addition, the XPS survey spectra of the PECs (data not shown) showed the presence of carbon, nitrogen, and oxygen but no chloride or sodium. This indicates that the PECs formed are composed predominantly of the polymer molecules. ATR-FTIR spectra of each of the polymers as well as the PECs formed at pH 3.8 are displayed in Figure 1.

Figure 1. FTIR characterization of insoluble matrix formed at pH 3.8 when mixing ALG or 40ADA with CHI: (a) ALG, (b) 40ADA, (c) CHI, (d) ALG/CHI[3.8], and (e) 40ADA/CHI[3.8].

The bands observed in the spectra of the polymers (Figure 1a− c) correlated with those previously reported.12,22 There was no difference evident in the FTIR spectra of PECs, formed from the same polymers in different pH solutions. The spectra of ALG/CHI[3.8] and 40ADA/CHI[3.8] (Figure 1d,e) show bands at 1600 cm−1 (COO− antisymmetric stretching), 1410 cm−1 (COO− symmetric stretching), 1320 cm−1 (C−O stretching), 1080 cm−1 (C−O−C symmetric stretching), 1030 cm−1 (C− O−C antisymmetric stretching), and 950 cm−1 (C−O stretching) originating from ALG or ADA.22,35 Bands at 1620 cm−1 (from antisymmetric NH3+ deformation, amide I), 1515 cm−1 (amide II stretching, NH bending, symmetric NH3+ deformation), 1150 cm−1 (antisymmetric C−O−C stretching, CN stretching), and 1060 cm−1 (CO stretching) originates from CHI.22,35 There is evidence of a carbonyl vibrational mode at 1735 cm−1 in the spectrum of 40ADA/CHI[3.8], Figure 1e. This band occurs at distinctly higher wavenumber than the carboxylic acid band of alginic acid (1710 cm−1)22 and correlates with an aldehyde band reported in previous studies to be present between 1730 and 1740 cm −1 . 12,36−38 Furthermore, this band is more pronounced in PECs with an increasing degree of oxidation of ADA (data not shown). 1809

DOI: 10.1021/acs.biomac.5b00383 Biomacromolecules 2015, 16, 1807−1817

Article

Biomacromolecules

Figure 2. XPS characterization of insoluble matrix formed at pH 3.8 when mixing ALG or 40ADA with CHI. XPS N 1s narrow scan spectra of (a) CHI, (b) ALG/CHI[3.8], and (c) 40ADA/CHI[3.8].

Figure 3. Confocal laser scanning microscopy images displaying 405, 488, 543, and 633 nm excitation and the overlay image of each: (a) ALG/ CHI[6.5] and (b) 40ADA/CHI[6.5].

silk fibroin where no change in the FTIR spectra was observed upon fixation with ADA (20% w/w).39 As pointed out in our previous study investigating the interactions between ALG and CHI, the use of XPS, not FTIR, is strongly recommended for the evaluation of the extent of electrostatic interactions in these polysaccharide PECs.22 The XPS N 1s narrow scans of CHI, ALG/CHI[3.8], and 40ADA/ CHI[3.8] are shown in Figure 2. Curve fitting to the spectra were performed by fixing the amide band (at 400.5 ± 0.1 eV) as constant at 9% in agreement with the degree of deacetylation for CHI22,30 and then refining the amine (at 399.4 ± 0.4 eV) and protonated amine (at 401.5 ± 0.5 eV) positions and intensities.45 The NH3+/Ntotal ratio differs in the samples CHI, ALG/CHI[3.8], and 40ADA/CHI[3.8] displaying values of 0.31, 0.51, and 0.47, respectively. Similar values are obtained for other PECs as included in Table 1. While the counterion in the CHI sample is a chloride ion, it is substituted by a carboxylic acid group from either ALG or 40ADA in the polyelectrolyte complexes (e.g., absence of Cl or Na in the XPS spectra as mentioned above). The NH3+/Ntotal ratios for the PECs fall in between those previously reported for chloride ion-free, strong ALG/CHI PECs (e.g., 0.55), and weak PECs (ratio of 0.10).22 This is supported by the observation that the ALG/CHI[6.5] and 40ADA/CHI[6.5] PECs of the current study showed no visible change when immersed in 10 × PBS for 1 week. We observed in the current study that the NH3+/Ntotal ratio is not significantly altered by pH of the solution in which the PECs

Relatively few studies report investigation of materials composed of CHI and ADA either as hydrogel matrices37,39−41 or in LbL assemblies.42,43 All these studies have inferred that imine formation occurs between the aldehyde of ADA and the amine of CHI from FTIR spectra. Some studies implied formation of an imine based on the disappearance of the aldehyde band at 1733 cm−1.37,41 However, we have shown that this band in ADA is very sensitive to the moisture content of the sample,12 and as such, it is not feasible to attribute changes in this band to infer imine formation. Other studies40,42 have assigned the appearance of a band at 1410 cm−1 to the imine stretching vibration; however, we and others have previously shown that a band at 1401−1416 cm −1 appears in polyelectrolyte complexes between ALG and CHI22,44 and that this band cannot be attributed to imine formation either. Evidence of the imine functionality (RHCNR) is displayed through the presence of a stretching vibrational mode in the FITR spectral range of 1674−1665 cm−1.35 Examples of assignment of bands to the imine have been reported in the PECs prepared by Mu et al.43 at 1654 cm−1 and in the gel prepared by Chen et al.37 at 1628 cm−1, both of which are outside the expected range. The amide I and NH3+ bands dominate the spectra of the PECs in this region as outlined above, and hence, based on our data including all reference samples, it is our opinion that identification of an imine functionality by FTIR is not feasible. This stance is supported by published data for a matrix consisting of 50% CHI and 50% 1810

DOI: 10.1021/acs.biomac.5b00383 Biomacromolecules 2015, 16, 1807−1817

Article

Biomacromolecules

deuterium exchange of the imine proton. However, since small molecules are suppressed during a diffusion-edited NMR experiment, the presence of resonances from the glucosamine molecules (clearly evident by the resonances at 2.90, 3.19, and 3.39 ppm Figure S3) provided evidence of a covalent reaction with 40ADA and hence imine formation. The reaction between the aldehyde groups of 40ADA and amine groups of glucosamine (molar equivalent amounts) was found to be substantially less than 100%. Considering the higher rate of diffusion of the small molecule glucosamine compared to CHI we predict that, although a lower concentration of imines would be present in the 40ADA/CHI PECs, this functionality is likely part of the associations in these materials (as evident by the salt immersion experiment); however, electrostatic interactions are likely to be more dominant (see discussion relating to XPS above). Overall, it can be concluded that the imine formation, which is likely to occur in the PECs containing ADA and CHI, cannot be confirmed or quantified by chemical characterization techniques commonly applied in this field (e.g., FTIR, XPS, or fluorescence). LbL Assembly of Alginate Dialdehyde and Chitosan on Silica Particles. Silica colloids were prepared by the Stöber process29 and were found to have an average diameter of 420 nm in good agreement with that predicted from the reagent ratios. For the LbL deposition on silica colloids, the effect of the type of ADA polymer was explored using XPS and zeta potential measurements investigating the (CHI[3.8]/20ADA)Ca and (CHI[3.8]/40ADA)Ca assemblies. The zeta potential of the silica particles was determined as −37 mV in agreement with previous studies.45 The zeta potential was observed to change significantly after addition of each layer as shown in Figure 4.

were synthesized (see data in Table 1 for PECs formed at pH 3.8 and 6.5) indicating that the electrostatic interaction between these weak polyelectrolytes is less responsive to pH than in previously prepared PECs between the strong polyelectrolyte dermatan sulfate and CHI.30 In our PEC system, a lower pH results in protonation of both the NH2 group of CHI (pKa 6.346) and the COO− group of ALG or ADA (pKa 3.4−3.747), while a higher pH decreases the amount of protonation. Evaluation of imine formation from the N 1s narrow scans is not feasible as the binding energy of the imine is not different to that of the amine, e.g., imines have been reported to have binding energies of 399.1,48 399.7,49 and 400 eV50 overlapping with the position of the amine peak at 399.4 ± 0.4 eV. There was no significant difference in the NH3+/Ntotal ratio of the different PEC samples, which leads to the conclusion that electrostatic interactions play a dominant role in complexes formed between either ALG or ADA and CHI. Researchers have proposed that fluorescence properties of CHI-based materials containing aldehydes including ADA in a LbL assembly is due to the n−π electronic transition of the imine group;42 however, their work did not include the pure polymers as control samples. To evaluate whether fluorescence properties can be used as an indication of the presence of imine groups, confocal microscopy of ALG, 40ADA, CHI, and the PECs was explored in the current study. The resulting confocal microscopy images, using excitation wavelengths of 405, 488, 543, and 633 nm, and overlay images are shown for the PECs in Figure 3 (confocal microscopy images for the three pure starting polymers are provided in Supporting Information, Figure S1). It was found that dried ALG, 40ADA, and CHI all separately exhibit fluorescence when excited at 405 and 488 nm. In addition, fluorescence of a lower intensity was observed when the polysaccharides were excited at 543 and 633 nm. These findings are in agreement with these polymers being semicrystalline.51,52 Confocal microscopy images of the freezedried ALG/CHI[6.5] and 40ADA/CHI[6.5] PECs likewise exhibited fluorescence on excitation at 405, 488, 543, and 633 nm (Figure 3). The PECs fluoresced strongly when excited at all wavelengths, but the highest intensity fluorescence was observed when excited at 405 and 488 nm, consistent with fluorescence of the pure polysaccharides. Based on these experiments and the absence of imine functionalities in the starting polymers as well as the ALG/CHI[6.5] PEC, we can conclude that confocal microscopy or other techniques measuring fluorescent properties of these types of complexes cannot be used to verify the presence of imine groups. Immersion of the ALG/CHI[6.5] and 40ADA/CHI[6.5] PECs in a saturated salt solution over a 1 week period revealed a difference in stability of these two PEC systems. While the ALG/CHI[6.5] PEC fully dissolved, the 40ADA/CHI[6.5] PEC remained solid during this process indicating that the interactions between the two biopolymers in this PEC were not solely electrostatic in nature and implies imine formation (Figure S2). In order to illustrate that imines do indeed form between the aldehyde groups of 40ADA and amines under aqueous solution conditions, the reaction with glucosamine (a model molecule for CHI) was evaluated by standard 1H NMR and diffusion-edited NMR (details are provided in the Supporting Information). Direct observation of imine functionalities could not be made as no new resonance that could be assigned to an imine resonance appeared in the 1H NMR spectra of the mixtures (when looking in the spectral region from 0 to 15 ppm, data not shown), which is likely due to

Figure 4. ζ-Potential values determined in 5 mM MES buffer (pH = 7.0; I = 10 mM) for Stöber silica particle samples (CHI[3.8]/20ADA)Ca (▲, full line) and (CHI[3.8]/40ADA)Ca (■,broken line). The lines are to guide the eye.

The addition of a CHI layer resulted in an increase in the zeta potential yielding values in the range of −16 to −13 mV for the two first layers, while the third CHI layer yielded particles with a zeta potential close to zero. These particles displayed a diameter of approximately 3 μm indicating a high degree of aggregation. The addition of each ADA layer resulted in a significant decrease in ζ-potential reaching values in the range of −33 to −37 mV, and these particles were observed to be colloidally stable. The lack of charge reversal evident upon CHI adsorption is attributed to the medium in which the zeta potential measurements were determined, which was a MES buffer with a pH of 7.0. The observed ζ-potential values for the CHI adsorbed silica particles are in agreement with the pKa 1811

DOI: 10.1021/acs.biomac.5b00383 Biomacromolecules 2015, 16, 1807−1817

Article

Biomacromolecules value for CHI being 6.3,46 and ζ-potential values for CHI at pH 4.0 and 5.5 reported as +17 and +13 mV, respectively.53 Many studies have used ζ-potential measurements of particle suspensions to monitor the buildup of alternating charged layers in LbL assemblies containing CHI; however, many do not report the solution conditions. Of those studies that do, values in the range of +15 to +23 mV have been found at pH of 5.0−5.6 in the second or higher CHI layer number (e.g., once substrate effects are not dominating),23,25−27 also in good agreement with the values reported for CHI. The ζ-potential data of the current study thus confirms effective buildup of LbL assemblies from these biopolymers. The dried silica particles were characterized by XPS after each layer deposition. The survey scans of the substrate revealed the presence of Si 2p and O 1s as well as minor C 1s contamination (data not shown). As a measure of the LbL assembly process the C 1s (arising primarily from the polymers) relative to Si 2p (arising from the substrate) was evaluated, and the result is shown in Figure 5A, which depicts

samples, the absence of the N 1s signal can be attributed to it being below the detection limit for XPS. Upon adsorption of each of the first four layers, relatively small increases in the C/Si atomic ratio were observed. However, in samples with five layers adsorbed the C/Si and N/Si atomic ratios increased dramatically compared to the sample with three and four layers adsorbed. This correlates with the observations of the zeta potential measurements. Only minor differences were observed between the (CHI[3.8]/20ADA)Ca and (CHI[3.8]/40ADA)Ca LbL assemblies, e.g., N/C atomic ratios for the five layer assemblies of 0.053 and 0.042, respectively, being similar based on an estimated error of 0.007, indicating comparable adsorption behavior for the two negatively charged polymers on this substrate. Overall the data obtained for the silica particles confirms the successful adsorption of each polyelectrolyte layer in the five-layered assemblies. LbL Assembly of Alginate Dialdehyde and Chitosan on Glass Slides. The five LbL assemblies (CHI[3.8]/ 20ADA) Ca , (CHI [3.8] /40ADA) Ca , (CHI [3.0] /20ADA) Ca , (CHI[3.0]/40ADA)Ca, and (CHI[3.0]/20ADA)Mg were investigated on the piranha treated glass slides as a substrate allowing evaluation of the effect of the type of ADA; the acidic solution used to dissolve CHI; and the nature of the supporting electrolyte added to both polyelectrolyte solutions. XPS was used to characterize the surface composition, and peaks for the elements N, Si, C, and O were observed in all samples, while Ca was observed in the unmodified glass slide as well as in some samples where CaCl2 was the supporting electrolyte. Atomic percentage ratios are displayed in Figure 6. A general observation from this data is that the first three layers resulted in only minor changes to the C/Si ratio (Figure 6A), while more pronounced increases were seen upon adsorption of the fourth and fifth layers. It was found that the composition of the solution that CHI was dissolved in only had a minor effect on the surface chemistry of the LbL assembly. As can be seen from Figure 6A, the most pronounced effect of the solution composition was in adsorption of the first layer. Adsorption of CHI from the acetic acid/acetate solution did not give a significant increase in the N 1s atomic percent, while adsorption of CHI from the citric acid buffer did (Figure 6B). The very low atomic percentages of N 1s detected when using CHI in acetate/acetic acid solution was accompanied by a minimal change in the C 1s component (Figure 6B). When the first CHI layer is adsorbed to the solid substrate (pH 3.0 or 3.8) the ionization of both the substrate and the polyelectrolyte is relatively high; however, charge shielding is expected due to the presence of sodium and citrate or acetate ions (ionic strengths of 0.20 and 0.22 M, respectively) in solution. While the type of ADA only had a minor effect on the LbL assembly on the silica particles, it had a large effect on the LbL assembly on the glass slide for the fourth and fifth layers. Regardless of the CHI solution composition the different ADA polymers resulted in very different C/Si atomic ratios with the 40ADA displaying a much larger value of 16−17 for the fourth layer compared to 5−7 for the 20ADA assemblies (Figure 6A). Notably, for this fourth layer the N/Si ratio increased significantly (from 0.2 to 0.7 or 0.9) for the 40ADA assemblies, while no significant change was seen for the 20ADA assemblies (Figure 6B). This observation must of course be seen in the context of all the LbL assemblies displaying the expected trend in the N/C ratio (shown in Figure 6C) of increasing with adsorption of every CHI layer and decreasing with every ADA layer. During adsorption of the ADA layers, this polymer will

Figure 5. XPS data ratio for Stöber silica particle samples (CHI[3.8]/ 20ADA)Ca (▲, full line) and (CHI[3.8]/40ADA)Ca (■, broken line). (A) C/Si atomic ratio; (B) N/Si atomic ratio. The lines are to guide the eye.

the ratio between the atomic ratios of C 1s relative to Si 2p. An increase in the C/Si ratio for each added polyelectrolyte layer was observed in agreement with buildup of the LbL assembly. In addition, the N 1s (arising exclusively from CHI) was used to evaluate the presence of this polymer (Figure 5B); however, it was not possible to detect an N 1s signal after adsorption of the first CHI layer, neither was it possible to detect a N 1s signal after the second ADA layer was adsorbed. Considering the very regular changes in zeta potential for these hydrated 1812

DOI: 10.1021/acs.biomac.5b00383 Biomacromolecules 2015, 16, 1807−1817

Article

Biomacromolecules

previously proposed.24 A final analysis of the XPS data for the LbL assemblies resulting from polyelectrolyte solutions containing CaCl2 involved the evaluation of the Ca-content with the Ca/Si data displayed in Figure 6D. In general, the Ca/ Si ratio is larger when ADA is the outer polyelectrolyte (layers 2 and 4), and this is most pronounced for the fifth layer. This can be related to the high affinity of Ca2+ for the carboxylate group. Upon adsorption of an ADA layer, it is proposed that the Ca2+ ions are extracted from the solution, while upon adsorption of a CHI layer, the acid used to dissolve this polyelectrolyte causes the Ca2+ ions to be extracted from the LbL assembly. The adsorption of CHI occurs at a pH where the ADA layer will be neutral, and in addition, charge shielding will take place. A final important observation from the data presented in Figure 6 is that the trend in Ca/Si ratio for the fourth layer (ADA) is the same as the trend in the N/Si ratio of the fifth layer (CHI) indicating that the amount of Ca2+ incorporated into the adsorbed ADA layer directly influences the amount of CHI subsequently adsorbed. In order to allow discrimination between purely ionic strength and specific binding effects in these assemblies, a (CHI[3.0]/20ADA)Mg LbL assembly was investigated. It was found by XPS survey scans that there were no Mg2+ ions incorporated into these LbL assemblies (data not shown). This observed difference in the behavior of the two divalent ions is similar to the published observation that Ca2+ causes gelation of ALG due to strong interaction with the carboxylate groups, while no gelling is observed in the presence of Mg2+.54 We can therefore conclude that the incorporation of Ca2+ in the LbL assemblies when ADA is the outer layer is caused by specific interactions between ADA and Ca2+. Evaluation of the XPS data found for the two-layer system that there was no significant difference between the (CHI[3.0]/20ADA)Mg and (CHI[3.0]/20ADA)Ca LbL assemblies with respect to the amount of ADA adsorbed; however, for the four-layer system the C/Si and N/Si ratios were significantly lower for the (CHI[3.0]/20ADA)Mg assembly compared to the (CHI[3.0]/ 20ADA)Ca assembly (C/Si of 3.6 versus 5.3 and N/Si of 0.13 versus 0.31, respectively). It thus appears that the film thickness is directly affected by the presence of the Ca2+ ions. The adsorption of the first three layers is not highly dependent on the solution composition or polymer composition and relatively small amounts of polyelectrolyte is adsorbed. However, large variations in layer thickness (assessed from C/ Si atomic ratio) are observed for the fourth and fifth layers and proposed structures of these layers are illustrated in Scheme 1. The thickness of the fourth layer is dependent on both the type of ADA polymer and the type of salt added (Ca2+ or Mg2+) with the thickness decreasing in the order of (CHI[3.0]/ 40ADA)Ca > (CHI[3.0]/20ADA)Ca > (CHI[3.0]/20ADA)Mg. In addition, Ca2+ is incorporated into this layer in larger amounts for the (CHI[3.0]/40ADA)Ca system. The increased N/Si ratio for the fourth (ADA) layer of the (CHI[3.0]/40ADA)Ca and (CHI[3.8]/40ADA)Ca assemblies indicates that CHI is most likely diffusing out from the LbL assembly toward the surface. Such rearrangements during the adsorption process have previously been reported for CHI/ALG multilayer formation,55 as well as for a range of other amines where it was found that partially charged polycations had high mobility allowing diffusion within the LbL assemblies.56 The adsorption of the fifth CHI layer is affected by the thickness and Ca2+ content in the preceding ADA layer with the same relative thickness resulting for this layer as that observed for the fourth layer. The

Figure 6. XPS atomic ratios for glass slide samples (CHI[3.8]/ 20ADA)Ca (▲, full line); (CHI[3.8]/40ADA)Ca (■, broken line); (CHI[3.0]/20ADA)Ca (◆, dot-broken line); and (CHI[3.0]/40ADA)Ca (●, dotted line). (A) C/Si atomic ratio; (B) N/Si atomic ratio; (C) N/C atomic ratio; (D) Ca/Si atomic ratio. The lines are to guide the eye.

carry a negative charge (pH 6.0), while CHI will be uncharged. This will result in predominantly H-bonding interactions as 1813

DOI: 10.1021/acs.biomac.5b00383 Biomacromolecules 2015, 16, 1807−1817

Article

Biomacromolecules

Scheme 1. Schematic Illustration of How the Chemical Differences between the 4th Layers of the LbL Assemblies Lead to Differences in Layer Thickness of the 5th Layer

assembly process.19,20,57 A study investigating LbL assemblies of polyelectrolytes around an agarose gel furthermore found the use of a weak (rather than strong) positively charged polyelectrolyte resulted in the highest encapsulation efficiency of bovine serum albumin.58 The type of negatively charged polyelectrolyte had less of an effect on the encapsulation efficiency. In the current study, the weak polyelectrolyte CHI was therefore used, and in addition, ADA was explored as the negatively charged polyelectrolyte in the LbL assembly processes around Ca-ALG hydrogels. The choice of ADA stems from the increased chain flexibility of this polymer due to cleavage of the C2−C3 bond upon oxidation rendering it nongelling in the presence of Ca2+ ions.9 Alginate hydrogels were prepared via Ca2+ cross-linking and the following LbL assemblies investigated: (CHI [3.8] / 20ADA)Ca, (CHI[3.0]/20ADA)Ca, (CHI[3.0]/40ADA)Ca, and

difference in the LbL assemblies incorporating the 20ADA and 40ADA polymers might be related to the difference in chain flexibility with the 40ADA having more conformational freedom thus enabling more loops forming upon adsorption which in turn enhances diffusion of CHI through a flexible polymer layer. LbL Assembly of Alginate Dialdehyde and Chitosan on Alginate Hydrogels. Build-up of LbL films around a calcium cross-linked alginate hydrogel (Ca-ALG) has been applied to enhance retention of encapsulated drugs to prolong their release and for cell encapsulation to limit diffusion of, e.g., immunogenic proteins. In studies using LbL assemblies to enhance retention of encapsulated drugs the polyelectrolyte solutions of positively charged CHI and the negatively charged dextran sulfate have included Ca2+ ions in order to prevent extraction of Ca2+ ions from the Ca-ALG core during the LbL 1814

DOI: 10.1021/acs.biomac.5b00383 Biomacromolecules 2015, 16, 1807−1817

Article

Biomacromolecules (CHI[3.0]/20ADA). Characterization of the surface chemistry, using XPS after each layer deposition, was carried out on dried gels. The XPS spectra of the Ca-ALG hydrogels showed the presence of C 1s, O 1s, and Ca 2p as well as a small amount of N 1s from contamination in the alginate (around 2 atom %). Only the original Ca-ALG gels displayed a Ca 2p atomic percent above 1%, and the evaluation of successful LbL assembly on the gels was therefore restricted to examining the N/C ratio as shown in Figure 7 which shows the data for

Figure 7. Change in surface chemistry evaluated by XPS of dried CaALG bead samples (CHI[3.8]/20ADA)Ca (▲, full line); (CHI[3.0]/ 20ADA)Ca (◆, dot-broken line); and (CHI[3.0]/40ADA)Ca (●, dotted line). The lines are to guide the eye. Figure 8. (A) Water uptake (Uw) and (B) swelling degree (Sw) of CaALG bead samples (CHI[3.0]/20ADA)Ca (◆, dot-broken line) and (CHI[3.0]/40ADA)Ca (●, dotted line). The lines are to guide the eye. n ≥ 5.

assemblies made in the presence of CaCl2 in the polyelectrolyte solutions. In general, in support of successful LbL assembly on the Ca-ALG beads, an increase in the N/C ratio was observed after each CHI layer adsorption while a decrease was observed after the ADA adsorption. This trend was most pronounced for the (CHI[3.0]/20-ADA)Ca assembly. Similar to what was observed for the silica particles, only a very minor change in the N/C ratio was observed upon adsorption of the first CHI layer when CHI was dissolved in the acetic acid/acetate solution. Similarly, buildup of CHI and dextran sulfate around an alginate microgel found that the XPS S/N ratio as well as the zeta potential did not stabilize to a regular change with each added layer until after the fourth layer had been applied.20 In that study, this was attributed to incomplete coverage of the polyelectrolytes during the initial layer adsorption processes. However, other studies using agarose58 or poly vinyl alcohol59 gels have reported polyelectrolyte diffusion into the gel matrix during LbL deposition in aqueous solution. Considering that in our study we see similar effects on silica particles and Ca-ALG gels, we attribute the initial low N/C ratio to incomplete coverage. In addition to evaluating the changes in surface chemistry of the dried Ca-ALG hydrogels upon LbL adsorption, the hydrated mass and the diameter of the Ca-ALG hydrogels coated in (CHI[3.0]/20ADA)Ca and (CHI[3.0]/40-ADA)Ca LbL assemblies were measured at each adsorption step and the water uptake (Uw) and swelling degree (Sw) used as an indication of changes to the Ca-ALG matrix. The data obtained are depicted in Figure 8 and show that there is good correlation observed between Uw and Sw for the systems investigated. It can be seen that upon adsorption of CHI, a decrease in the hydrated mass (decrease in Uw) and swelling degree (decrease in Sw) resulted. In a control experiment a similar effect was observed upon submerging Ca-ALG hydrogels in the citric acid

buffer without added CHI or CaCl2 where an Uw of −20% was found. These values correlate with the bulk Ca2+ content in the beads after immersion in the CHI citric acid buffer containing CaCl2 (64% retained) compared to immersion in the buffer alone (52% retained). Upon immersion of the Ca-ALG gels in the CHI solutions a number of factors will affect the swelling/syneresis outcome. First, it is affected by the pH of the respective solutions. The pKa value for ALG is 3.4−3.6, and it has been reported that alginic acid forms below pH 3.6.47 Considering that an alginic acid gel is less hydrated than a Ca-ALG gel, the reduction in Uw upon immersion in the pH 3 citric acid buffer could be linked to formation of alginic acid at least at the gel’s surface. This would be expected to be accompanied by extraction of Ca2+ ions from the gel. Second, as pointed out above when describing the LbL assemblies onto glass slides, the acids used to dissolve CHI have a high affinity for Ca2+ ions. This was observed directly by determining the retention of Ca2+ crosslinker by ICP after immersion in the respective buffers. It should be noted that the alginate gel beads will have a polymer and Ca2+ ion gradient with higher concentrations at the perimeter,60 and thus, the magnitude of the reduction in crosslinker ions may be attributed to a surface effect. Third, the different ionic strengths of the Ca-ALG gels and the external solution also potentially affect the hydration of the gel. The remaining LbL assembly process for the (CHI[3.0]/ 20ADA)Ca and (CHI[3.0]/40ADA)Ca assemblies caused relatively less pronounced effects with significantly smaller changes in gel dimensions and water content regardless of the ADA polymer with increased values of Uw and Sw with addition of an 1815

DOI: 10.1021/acs.biomac.5b00383 Biomacromolecules 2015, 16, 1807−1817

Article

Biomacromolecules

confocal microscopy. R.A. would like to thank the Australian government for awarding an Australian Postgraduate Award and the Queensland Government for Ph.D. Scholarship Funding.

ADA layer (layers 2 and 4) and decreased values upon adsorption of a CHI layer (layers 3 and 5). A study using LbL of CHI and dextran sulfate around alginate microgels used polyelectrolyte solutions both at pH 5.6 and high CaCl2 concentrations;20 however, they did not describe whether the gel volume remained constant during the process. In the use of ADA as the anionic polyelectrolyte in the current study, the maximum concentration of Ca2+ that can be used was found to be 0.01 M as, despite reports of the nongelling properties of ADA,9 a weakly assembled hydrogel structure was observed at higher Ca2+ concentration. However, in our study, the benefit of adding a relatively low concentration of CaCl2 (0.01 M) to the polyelectrolyte solutions is clear when considering that CaALG beads with five layers of the (CHI[3.0]/20ADA) assembly applied resulted in a Uw value of 83% compared to values close to zero with the presence of CaCl2 (Figure 8). Overall this data demonstrates that successful LbL assembly around a Ca-ALG hydrogel requires optimization in order to avoid large dimensional and water content changes during the LbL assembly process.



(1) Grant, G. T.; Morris, E. R.; Rees, D. A.; Smith, P. J. C.; Thom, D. FEBS Lett. 1973, 32, 195. (2) Draget, K. I.; Østgaard, K.; Smidsrød, O. Appl. Microbiol. Biotechnol. 1989, 31, 79. (3) Lee, K. Y.; Mooney, D. J. Prog. Polym. Sci. 2012, 37, 106. (4) Pawar, S. N.; Edgar, K. J. Biomaterials 2012, 33, 3279. (5) Al-Shamkhani, A.; Duncan, R. J. Bioactive Compatible Polym. 1995, 10, 4. (6) Bouhadir, K. H.; Lee, K. Y.; Alsberg, E.; Damm, K. L.; Anderson, K. W.; Mooney, D. J. Biotechnol. Prog. 2001, 17, 945. (7) Painter, T.; Larsen, B. Acta Chem. Scand. 1970, 24, 813. (8) Smidsrød, O.; Painter, T. Carbohydr. Res. 1973, 26, 125. (9) Gomez, C. G.; Rinaudo, M.; Villar, M. A. Carbohydrate Polym. 2007, 67, 296. (10) Bouhadir, K. H.; Alsberg, E.; Mooney, D. J. Biomaterials 2001, 22, 2625. (11) Lee, K. Y.; Bouhadir, K. H.; Mooney, D. J. Biomaterials 2004, 25, 2461. (12) Jejurikar, A.; Seow, X. T.; Lawrie, G.; Martin, D.; Jayakrishnan, A.; Grøndahl, L. J. Mater. Chem. 2012, 22, 9751. (13) Balakrishnan, B.; Jayakrishnan, A. Biomaterials 2005, 26, 3941. (14) Xu, Y.; Li, Li.; Yu, X.; Gu, Z.; Zhang, X. Carbohydr. Polym. 2012, 87, 1589. (15) Crozier, T.; Boudou, T.; Picart, C. Curr. Opin. Colloid Interface Sci. 2010, 15, 417. (16) Almodovar, J.; Place, L. W.; Gogolski, J.; Erickson, K.; Kipper, M. J. Biomacromolecules 2011, 12, 2755. (17) Boddohi, S.; Killingsworth, C. E.; Kipper, M. J. Biomacromolecules 2008, 9, 2021. (18) Lundin, M.; Blomberg, E.; Tilton, R. D. Langmuir 2010, 26, 3242. (19) Matsusaki, M.; Sakaguchi, H.; Serizawa, T.; Akashi, M. J. Biomater. Sci. Polym. Ed. 2007, 18, 775. (20) Zuo, Q.; Lu, J.; Hong, A.; Zhong, D.; Xie, S.; Liu, Q.; Huang, Y.; Shi, Y.; He, L.; Xue, W. Biomed. Mater. 2012, 7, 035012. (21) Richert, L.; Lavalle, P.; Payan, E.; Shu, X. Z.; Prestwich, G. D.; Stoltz, J.-F.; Schaaf, P.; Voegel, J.-C.; Picart, C. Langmuir 2004, 20, 448. (22) Lawrie, G.; Keen, I.; Drew, B.; Chandler-Temple, A.; Rintoul, L.; Fredericks, P.; Grøndahl, L. Biomacromolecules 2007, 8, 2533. (23) Mihai, M.; Schwarz, S.; Janke, A.; Ghiorghita, C. A.; Dragan, E. S. J. Polym. Res. 2013, 20, 89. (24) Cuomo, F.; Lopez, F.; Ceglie, A.; Maiuro, L.; Miguel, M. G.; Lindman, B. Soft Matter 2012, 8, 4415. (25) Ye, S.; Wang, C.; Liu, X.; Tong, Z. J. Controlled Release 2005, 106, 319. (26) Zhao, Q.; Mao, Z.; Gao, C.; Shen, J. J. Biomater. Sci. Polymer Ed. 2006, 17, 997. (27) Wang, C.; Ye, W.; Zheng, Y.; Liu, X.; Tong, Z. Int. J. Pharm. 2007, 338, 165. (28) de Villers, M. M.; Otto, D. P.; Strydom, S. J.; Lvov, Y. M. Adv. Drug Delivery Rev. 2011, 63, 701. (29) Bogush, G. H.; Tracy, M. A.; Zukoski, C. F. J. Non-Cryst. Solid 1988, 104, 95. (30) Wen, Y.; Grøndahl, L.; Gallego, M. R.; Jorgensen, L.; Møller, E. H.; Nielsen, H. M. Biomacromolecules 2012, 13, 905. (31) Beamson, G.; Briggs, D. High Resolution XPS of Organic Polymers The Scienta ESCA300 Database; John Wiley & Sons Ltd: West Sussex, 1992. (32) Elzatahry, A. A.; Eldin, M. S.; Soliman, E. A.; Hassan, E. A. J. Appl. Polym. Sci. 2009, 111, 2452.



CONCLUSION The current study has demonstrated the difficulties with providing direct evidence of imine formation between the amine group of CHI and the aldehyde group of ADA using techniques commonly applied. Indirect evidence was obtained from evaluation of the stability of the 40ADA/CHI[6.5] PEC in a salt solution where it remained, while the control ALG/CHI[6.5] PEC dissolved. The LbL assemblies of the biopolymers showed interesting features, which depended on the presence of calcium ions in the polyelectrolyte solution and on the degree of oxidation of ADA. It was observed that when applying the LbL assembly to alginate hydrogels, significant swelling and syneresis occurred, and such evaluations are clearly important if aiming to use LbL to control drug diffusion from such gels.



ASSOCIATED CONTENT

S Supporting Information *

Experimental details for aqueous SEC; determination of degree of oxidation of ADA; NMR. Results for chemical characterization of chitosan; chemical characterization of ALG and ADA; confocal laser scanning microscopy images for control samples ALG, 40ADA, and CHI; photographic images of PECs immersed in saturated salt solution; diffusion-edited NMR spectra. The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/ acs.biomac.5b00383.



REFERENCES

AUTHOR INFORMATION

Corresponding Author

*Phone: +61 7 336 53671. Fax: +61 7 336 54299. E-mail: l. [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS The authors would like to acknowledge the following student who recorded data for LbL assembly on alginate gels in acetate buffer, Melissa Patricia Foong Shi Shan. Mr. George Blazak is acknowledged for the elemental microanalysis data. Dr. Barry Wood is acknowledged for assistance with XPS, Dr. Tri Le for assistance with NMR, and Dr. Steven Mason for assistance with 1816

DOI: 10.1021/acs.biomac.5b00383 Biomacromolecules 2015, 16, 1807−1817

Article

Biomacromolecules (33) Tan, K. L.; Tan, B. T. G.; Kang, E. T.; Neoh, K. G. Phys. Rev. B 1989, 39, 8070. (34) Kang, E. T.; Neoh, K. G.; Khor, S. H.; Tan, K. L.; Tan, T. G. J. Chem. Soc., Chem. Commun. 1989, 695. (35) Lin-Vien, D.; Colthup, N. B.; Fateley, W. G.; Grasselli, J. G. The Handbook of Infrared and Raman Characteristic Frequencies in Organic Molecules; Academic Press: London, 1991. (36) Le-Tien, C.; Millette, M.; Lacroix, M.; Mateescu, M.-A. Biotechnol. Appl. Biochem. 2004, 39, 189. (37) Chen, f.; Tian, M.; Zhang, D.; Wang, J.; Wang, Q.; Yu, X. Mater. Sci. Eng., C 2012, 32, 310. (38) Bouhadir, K. H.; Hausman, D. S.; Mooney, D. J. Polymer 1999, 40, 3575. (39) Gu, Z.; Xie, H.; Huang, C.; Li, L.; Yu, X. Int. J. Biol. Macromol. 2013, 58, 121. (40) Wang, J.; Fu, W.; Zhang, D.; Yu, X.; Li, J.; Wan, C. Carbohydrate Pol 2010, 79, 705. (41) Vieira, E. F. S.; Cestari, A. R.; Airoldi, C.; Loh, W. Biomacromolecules 2008, 9, 1195. (42) Jia, Y.; Fei, J.; Cui, Y.; Yang, Y.; Gao, L.; Li, J. Chem. Commun. 2011, 47, 1175. (43) Mu, B.; Lu, C.; Liu, P. Colloid Surf. B: Biointerfaces 2011, 82, 385. (44) Alves, N. M.; Picart, C.; Mano, J. F. Macromol. Biosci. 2009, 9, 776. (45) Lawrie, G. A.; Grøndahl, L.; Battersby, B. J.; Keen, I.; Lorentzen, M.; Surawski, P.; Trau, M. Langmuir 2006, 22, 497. (46) Yalpani, M.; Hall, L. D. Macromolecules 1985, 17, 272. (47) Haug, A. Norwegian Institute of Seaweed Research Report No. 30, 1964. (48) Gengenbach, T. R.; Chatelier, R. C.; Griesser, H. J. Surf. Interface Anal. 1996, 24, 611. (49) Fally, F.; Doneux, C.; Riga, J.; Verbist, J. J. J. Appl. Polym. Sci. 1995, 56, 597. (50) Wilson, D. J.; Williams, R. L.; Pond, R. C. Surf. Interface Anal. 2001, 31, 397. (51) Mogilevskaya, E. L.; Akopova, T. A.; Zelenetskii, A. N.; Ozerin, A. N. Polymer Science, Ser. A 2006, 48, 116. (52) Li, L.; Fang, Y.; Vreeker, R.; Appelqvist, I.; Mendes, E. Biomacromolecules 2007, 8, 464. (53) Costa, R. R.; Testera, A. M.; Arias, F. J.; Rodriguez-Cabello, J. C.; Mano, J. F. J. Phys. Chem. B 2013, 117, 6839. (54) Smidsrød, O.; Haug, A. Acta Chem. Scand. 1968, 22, 1989. (55) Maustad, G.; Mørch, Y. A.; Bausch, A. R.; Stokke, B. T. Carbohydr. Polym. 2008, 71, 672. (56) Zacharia, N. S.; Modestino, M.; Hammond, P. T. Macromolecules 2007, 40, 9523. (57) Anal, A. K.; Stevens, W. F. Int. J. Pharm. 2005, 290, 45. (58) Mak, W. C.; Cheung, K. Y.; Trau, D. Chem. Mater. 2008, 20, 5475. (59) Serizawa, T.; Sakaguchi, H.; Matsusake, M.; Akashi, M. J. Polym. Sci., Part A: Polym. Chem. 2005, 43, 1062. (60) Skjåk-Bræk, G.; Grasdalen, H.; Smidsrød, O. Carbohydr. Polym. 1989, 10, 31.

1817

DOI: 10.1021/acs.biomac.5b00383 Biomacromolecules 2015, 16, 1807−1817