Alternating bioactivity of multilayer thin films assembled from charged derivatives of chitosan

Journal of Colloid and Interface Science 316 (2007) 331–343 www.elsevier.com/locate/jcis

Alternating bioactivity of multilayer thin films assembled from charged derivatives of chitosan Somruethai Channasanon a , Wilaiporn Graisuwan a , Suda Kiatkamjornwong b , Voravee P. Hoven c,∗ a Program of Petrochemistry and Polymer Science, Faculty of Science, Chulalongkorn University, Phayathai Road, Pathumwan, Bangkok 10330, Thailand b Department of Imaging and Printing Technology, Faculty of Science, Chulalongkorn University, Phayathai Road, Pathumwan, Bangkok 10330, Thailand c Organic Synthesis Research Unit, Department of Chemistry, Faculty of Science, Chulalongkorn University, Phayathai Road, Pathumwan,

Bangkok 10330, Thailand Received 26 January 2007; accepted 26 July 2007 Available online 6 August 2007

Abstract Charged derivatives of chitosan, N -sulfofurfuryl chitosan (SFC) and N-[(2-hydroxyl-3-trimethylammonium)propyl]chitosan chloride (HTACC) were prepared by reductive alkylation of amino groups of chitosan (CHI) using 5-formyl-2-furansulfonic acid, sodium salt (FFSA) as a reagent and ring opening of glycidyltrimethylammonium chloride (GTMAC) by amino groups of chitosan, respectively. The chemical structures of the charged derivatives were verified by 1 H NMR and FTIR analyses. Multilayer assembly of SFC, HTACC, CHI and the selected oppositely charged polyelectrolytes was monitored by a quartz crystal microbalance (QCM). Stratification of the multilayer film fabricated on plasma-treated poly(ethylene terephthalate) (treated PET) substrate was demonstrated by water contact angle data. The coverage of the assembled films was characterized by AFM and ATR-FTIR analyses. The bioactivity of the deposited multilayer film on the treated PET substrate was tested against selected proteins having a distinctive size and charge. This research strongly suggests that both SFC and HTACC are potential candidates for altering the surface bioactivity of materials. © 2007 Elsevier Inc. All rights reserved. Keywords: Chitosan; Charged derivative; Layer-by-layer adsorption; Multilayer film; Polyelectrolyte; Protein adsorption

1. Introduction The performance of biomedical devices and biomaterials depends greatly upon the surface properties of the materials, since it is the surface of the materials that first comes into contact with their biological surroundings. Several approaches have been applied for the surface modification of materials to suit specific biomedical applications. Layer-by-layer (LBL) adsorption of oppositely charged polyelectrolytes is an efficient method of fabricating biocompatible polyion complex multilayer film on substrate [1]. This method has been developed over the past several years as a powerful yet simple strategy to engineer surfaces with specific properties. The process, driven by electrostatic interactions, involves sequentially dipping a charged substrate into dilute aqueous solutions of oppositely charged polyelec* Corresponding author. Fax: +66 2218 7598.

E-mail address: [email protected] (V.P. Hoven). 0021-9797/$ – see front matter © 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.jcis.2007.07.071

trolytes and allowing the polymer to adsorb and reverse the charge of the substrate surface. This aqueous-based technique is capable of varying film composition while providing enormous design flexibility. This versatile technique offers the benefit of solvent-free processing as well as the ability to conformably coat all available surfaces of virtually any material, irrespective of shape or size, with uniform ultrathin films of precisely controlled thickness. Chitosan is a partially deacetylated form of chitin, a natural substance found abundantly in the exoskeletons of insects, the shells of crustaceans, and fungal cell walls. In view of its biocompatibility, healing capabilities, and positively charged character, chitosan has often been selected as the primary choice for constructing multilayer films for biomedical applications. Alternating bioactivity against human blood and fibroblast adhesion was accomplished by chitosan/dextran sulfate composite thin films [2,3]. However, that was not the case for chitosan/heparin systems, suggesting that the polymer

332

S. Channasanon et al. / Journal of Colloid and Interface Science 316 (2007) 331–343

species and/or the assembly conditions are key factors in realizing the alternating bioactivities [4]. The same system was also used to generate antiadhesive and antibacterial coatings on aminolyzed poly(ethylene terephthalate) [5]. The deposition of self-assembled nanocoatings of chitosan and hyaluronan onto damaged arteries was described as a successful means not only to protect a damaged artery against thrombogenesis during the revascularization procedure, but also to control the healing processes by incorporating bioactive molecules within the multilayer films [6]. The same multilayer system embedded with isolated islets was found to be resistant to chondrocytes and bacteria [7]. Cytocompatibility of aminolyzed poly(L-lactic acid) [8], hydrolyzed poly(ethylene terephthalate) [9], and titanium [10] was improved by multilayer film of chitosan/poly(styrene sulfonate), chitosan/chondroitin sulfate, and chitosan/gelatin, respectively. Chitosan-based multilayer systems are also applicable for biosensor applications. A multilayer thin film of chitosan and glucose oxidase was applied as a coating on a Pt electrode in order to detect glucose [11]. Positively and negatively charged platinum nanoparticles blended with carbon nanotubes were prepared by mixing with chitosan and poly(styrene sulfonate), respectively, and then employed to fabricate multilayer films on a gold electrode and quartz glass slide. After the immobilization of the cholesterol oxidase onto the electrode, a constructed biosensor responded sensitively to cholesterol [12]. Although several research works have reported the formation of polyion complex multilayer film between chitosan and a number of anionic polyelectrolytes, none of these have mentioned charged derivatives of chitosan, whose biological properties are significantly different from those of chitosan. The positive charges of chitosan attract negatively charged plasma proteins leading to platelet adhesion and activation followed by thrombus formation and blood coagulation [13,14], whereas N -sulfofurfuryl chitosan [15] and sulfated chitosan [16] possess antithrombogenic properties, similar to what is observed for heparin. Quaternary ammonium chitosan has attracted considerable interest because of its improved aqueous solubility and antimicrobial activity over a broader pH range in comparison with native chitosan [17–19]. Antimicrobial activity is believed to originate from the ability of cationic species to bind with sialic acid in phospholipids, consequently restraining the movement of microbiological substances [20]. Oligomeric chitosan can also penetrate into the cells of microorganisms and prevent the growth of cells by prohibiting the transformation of DNA into RNA [21]. In particular, N -[(2-hydroxyl-3-trimethylammonium)propyl]chitosan chloride (HTACC) has been introduced as an effective antimicrobial agent for a number of fibers/fabrics such as cotton and cellulose [22–25]. In this study, we focus our attention on assembling polyion complex thin films from chitosan and its charged derivatives, SFC, and HTACC. Formation of polyion complex multilayer films from three pairs of oppositely charged polyelectrolytes is explored: chitosan (CHI) and poly(sodium styrene sulfonate) (PSS), poly(allylamine hydrochloride) (PAH) and N -sulfofurfuryl chitosan (SFC), N -[(2-hydroxyl-3-trimethylammonium)propyl]chitosan chloride (HTACC) and poly(acryl-

ic acid) (PAA). Contact angle analysis and a quartz crystal microbalance (QCM) are used as tools to follow the assembly process. The biological responses of all polyion complex multilayer films are tested by protein adsorption studies. We hypothesize that an alternate response can be achieved as long as each layer is thick enough and the overall biological response depends on the outermost layer. This study should provide fundamental information that can lead to the further development of chitosan derivatives for advanced applications in nanotechnology and biotechnology. 2. Materials and methods 2.1. Materials Chitosan flakes (95% DAC, Mw = 100,000) were obtained from Seafresh Chitosan (Lab) Co., Ltd. (Thailand). 5-Formyl2-furan-sulfonic acid, sodium salt (FFSA), poly(sodium styrene sulfonate) (PSS), Mw = 70,000, poly(allylamine hydrochloride) (PAH), Mw = 70,000, bovine serum albumin (BSA), fibrinogen (FIB), lysozyme (LYZ), γ -globulin (GLB), and phosphate buffer saline (PBS) were supplied by Aldrich (USA). Triethanolamine, sodium borohydride (NaBH4 ), glycidyltrimethylammonium chloride (GTMAC), sodium hydroxide (NaOH), conc. hydrochloric acid, poly(acrylic acid) (PAA), Mw = 60,000, and sodium dodecyl sulfate (SDS) were purchased from Fluka (Switzerland). A bicinchoninic acid assay kit (QuantiPro™ BCA assay) was purchased from Sigma Chemical Co. (USA). Methanol, ethanol, acetone, glacial acetic acid, and sodium chloride were of reagent grade and purchased from Merck (Germany). Ultra-pure distilled water was obtained after purification by a Nanopure system (Barnstead Thermolyne Corporation, USA). Plasma-treated poly(ethylene terephthalate) films (i.d. = 14 mm) were purchased from Wako Pure Chemical Industry, Ltd. (Japan). 2.2. Nuclear magnetic resonance spectroscopy (NMR) The 1 H NMR spectra were recorded in D2 O using a Varian model Mercury-400 nuclear magnetic resonance spectrometer (USA) operating at 400 MHz. Chemical shifts (δ) were reported in parts per million (ppm) relative to tetramethylsilane (TMS) or using the residual protonated solvent signal as a reference. 2.3. Fourier transform-Infrared spectroscopy (FTIR) The FT-IR spectra were recorded with a FTIR spectrometer (Perkin–Elmer, USA), model system 2000, with 32 scans at resolution 4 cm−1 . Data at frequencies of 400–4000 cm−1 were collected by using a TGS detector. 2.4. Quartz crystal microbalance (QCM) The apparatus and the crystals used for QCM measurements were obtained from Maxtek, Inc. (USA). An AT-cut quartz crystal with a resonance frequency of 5 MHz, model SC-501-1

S. Channasanon et al. / Journal of Colloid and Interface Science 316 (2007) 331–343

was used. The plating crystal (1 inch in diameter) was covered by evaporated gold on both faces. The frequency shift of the gold-coated QCM plate was monitored by a Maxtek plating monitor (model PM-710) coupled with a MPS-550 sensor probe. The Data-Log Software (Maxtek) was used for data acquisition and monitoring. The gold-coated QCM plate was cleaned by soaking in piranha solution (3:1, H2 SO4 :30% H2 O2 ) for 5 min, rinsing thoroughly with nanopure water and drying with a light stream of nitrogen gas prior to use. The surface modification of the gold-coated QCM plate was carried out by immersion in an ethanolic solution containing 1:1 (v/v) of 10 mM 11-mercaptoundecanoic acid and 10 mM 11-mercapto1-undecanol at ambient temperature. After 24 h, the plate was rinsed with ethanol and dried with a light stream of nitrogen gas. For the measurements in the in situ mode, one side of the gold-coated QCM plate was kept in permanent contact with a designated polyelectrolyte solution and the frequency change was recorded continuously. For the measurements in the conventional mode, one side of the gold-coated QCM plate was covered with a designated polyelectrolyte solution for a desired period of time, rinsed with copious amounts of nanopure water, dried with a light stream of nitrogen gas and then the resonance frequency was measured in air. In the case of multilayer assembly, rinsing with copious amounts of nanopure water and drying by a light stream of nitrogen gas were applied after each step of deposition. 2.5. Contact angle measurements A contact angle goniometer model 100-00 equipped with a Gilmont syringe and a 24-gauge flat-tipped needle (Ramé-Hart, Inc., USA), was used for the determination of water contact angles. A droplet of nanopure water was placed on the tested surface by bringing the surface into contact with a droplet suspended from a needle on the syringe. The measurements were carried out in air at room temperature. The reported angle is an average of 5 measurements on different areas of each sample. 2.6. Attenuated total reflectance-Fourier transform infrared spectroscopy (ATR-FTIR) All spectra were collected at a resolution of 4 cm−1 and for 128 scans using a Nicolet Magna 750 FTIR spectrometer equipped with a liquid-nitrogen-cooled mercury–cadmium– telluride (MCT) detector. A single attenuated total reflection accessory with 45◦ germanium (Ge) IRE (Spectra Tech, USA) and a variable angle reflection accessory (Seagull™, Harrick Scientific, USA) with a hemispherical Ge IRE were employed for all ATR spectral acquisitions. 2.7. Atomic force microscopy (AFM) AFM images were recorded with a Scanning Probe Microscope, model NanoScope® IV, Veeco, USA. Measurements were performed in air at ambient temperature using tapping mode. Silicon nitride tips with a resonance frequency of 267– 298 kHz and a spring constant of 20–80 N/m were used.

333

2.8. Synthesis of N -sulfofurfuryl chitosan (SFC) SFC was synthesized according to a method modified from that of Amiji [15]. Chitosan flakes (0.20 g, 1 equiv. of NH2 ) were dissolved in 1% (v/v) aqueous acetic acid (10 mL) to prepare a 2.0% (w/v) solution. Methanol (10 mL) containing 1.0% (w/v) triethanolamine was slowly added to the chitosan solution. The mixture was stirred for 6 h at room temperature. FFSA (0.46 g, 2 equiv.) was slowly added to the chitosan slurry. The reaction was allowed to proceed for 18 h at room temperature. As the reaction continued, the Schiff base thus formed was reduced by slow addition of NaBH4 (6 equiv.). After reduction for 6 h, the Schiff base slowly dissolved to form a viscous solution. SFC was precipitated in methanol and washed extensively with methanol and acetone to remove the unreacted FFSA. The polymer was dried at room temperature in a vacuum oven. Then it was kept in a desiccator and milled to produce fine particles. 2.9. Synthesis of N -[(2-hydroxyl-3-trimethylammonium)propyl]chitosan chloride (HTACC) HTACC was synthesized according to a method modified from that of Seong et al. [23]. Chitosan (0.4 g, 1 equiv. of NH2 ) was dissolved in 1% acetic acid to prepare a 2.0% (w/v) chitosan solution. GTMAC (1.4 g, 4 equiv.) was added. The reaction was performed at 70 ◦ C for 24 h. After the reaction, the solution was poured into an acetone/ethanol (50:50, v/v) mixture to obtain the precipitate. The precipitate was filtered, washed thoroughly with acetone, dried under vacuum at room temperature and kept in a desiccator. 2.10. Pre-treatment of plasma-treated poly(ethylene terephthalate) (treated PET) substrate Treated PET substrates were soaked in sodium hydroxide solution (1 M) at 60 ◦ C for 1 h. They were then immersed in hydrochloric acid (0.1 M) for 10 min at room temperature. Finally, the substrates were rinsed thoroughly with nanopure water and air-dried at room temperature. 2.11. Polyelectrolyte multilayer assembly The alternating layers were assembled by sequentially dipping the treated PET substrates in a polycation solution (1 mg/mL of CHI, 2 mg/mL of PAH, or 3 mg/mL of HTACC) and a polyanion solution (2 mg/mL of PSS, 2 mg/mL of SFC, or 2 mg/mL of PAA) for 30 min interval. Then 1 M NaCl was added into the polymer solution. All polyelectrolytes were not buffered. The pH of the solution was adjusted by using HCl (aq) or NaOH (aq). Three pairs of polyelectrolyte self-assemblies were fabricated at pH 4 for CHI–PSS, at pH 8 for PAH–SFC and at pH 7 for HTACC–PAA. The substrates were rinsed thoroughly with nanopure water between each dipping and after the final adsorption. After the desired number of layers was deposited, the substrates were blow-dried by a light stream of nitrogen gas.

334

S. Channasanon et al. / Journal of Colloid and Interface Science 316 (2007) 331–343

2.12. Protocol for protein adsorption test The deposited multilayer films on treated PET substrates were placed into a 24-well plate containing nanopure water overnight to reach an equilibrium hydration. Each substrate was removed from the nanopure water and suspended in a well containing 2.0 mL of protein (BSA, LYZ, FIB or GLB) solution then incubated at 37 ◦ C for 3 h. Three samples were analyzed for each condition. The samples were removed from the protein solution and rinsed thoroughly with PBS (2×) to remove any loosely attached protein. The adsorbed protein on the sample surface was detached by soaking each sample in 2.0 mL of 1% aqueous solution of SDS for 30 min. A protein analysis kit based on the BCA method was used to determine the concentration of the protein dissolved in the SDS solution. The 100 µL (0.1 mL) of SDS solution that soaked each sample was added into a well of a 96-well plate. Then 100 µL of BCA working solution was added into each well. The well-plate was incubated at 37 ◦ C for 2 h. The absorbance of the solution was measured at 562 nm by UV–vis spectroscopy (Microplate reader, Model EL340, Bio-Tek™ Instruments Inc., USA). The amount of protein adsorbed on the samples was calculated from the protein concentration in the SDS solution. The data are expressed as mean ± standard deviation (S.D.). 3. Results and discussion 3.1. Synthesis and characterization of charged derivatives of chitosan In comparison with CHI, 1 H NMR and FTIR spectra of SFC and HTACC are illustrated in Figs. 1 and 2, respectively. SFC

exhibited 1 H NMR signals of two furan protons at 6.3 and 6.7 ppm, indicating the attachment of the heterocyclic component. On the basis of the 1 H NMR calculation, the maximum %DS of amino groups of chitosan by N -sulfofurfuryl groups of 60% was obtained using 2 equivalents of FFSA. The appearance of a peak at 1233 cm−1 in the FTIR spectrum of SFC signified the presence of the sulfonic acid functionality. Signals corresponding to the protons of CH2 (2.6 ppm), CH (4.4 ppm), and CH3 (3.1 ppm) appearing in the 1 H NMR spectrum of HTACC together with their peak area suggested that both hydroxyl and amino groups reacted with GTMAC and the quaternary ammonium groups of N+ (CH3 )3 were incorporated. The %DS of GTMAC on chitosan was calculated from the ratio between the peak integration of the protons from the quaternary ammonium group at 3.1 ppm for GTMAC and the methyl protons of the acetamide group at 2.1 ppm for chitosan. The condition employing 4 equivalents of GTMAC yielded the highest %DS of ∼96%. According to FTIR analysis, the formation of HTACC was also verified by the decrement of the N–H scissoring signal at 1599 cm−1 and the increment of the C–H bending of the methyl group at 1482 cm−1 . The solubility test was performed according to the method previously described [26]. A solid sample of a charged derivative of chitosan (60 mg) was dissolved in water (20 mL). The pH of the solution was adjusted with 0.5% (w/v) aqueous HCl and NaOH. SFC exhibited solubility only in the alkaline range (pH 8–13), and insolubility in the acidic (pH 2–6) and neutral (pH 7) ranges. In contrast, HTACC with quaternary ammonium groups showed solubility over the entire pH range.

Fig. 1. 1 H NMR spectra of CHI, HTACC, and SFC.

S. Channasanon et al. / Journal of Colloid and Interface Science 316 (2007) 331–343

335

Fig. 2. FTIR spectra of CHI, HTACC, and SFC.

Fig. 3. In situ monitoring of frequency shift (Hz) as a function of time of SFC having different concentrations: (a) 3, (b) 2, and (c) 1 mg/mL, without NaCl at pH 8.

3.2. Adsorption of charged derivatives of chitosan To determine the adsorption time at which the adsorption of each polyelectrolyte has reached equilibrium, in situ monitoring of adsorption on a clean gold-coated QCM plate was conducted using QCM measurement. Fig. 3 shows the frequency shift of the QCM plate due to SFC adsorption. A greater frequency shift was observed when a higher concentration of SFC was used. A similar trend was also found for the adsorption of HTACC (data not shown). Such a trend may stem from the fact that the higher polyelectrolyte concentration generally promotes physical adsorption in a multilayer fashion, which does not persist after thorough rinsing. As a result, the magnitude of the frequency change shown in Fig. 3 should not correspond to the actual quantity of polyelectrolyte adsorbed on the

gold-coated QCM plate, because rinsing was not applied in the experiment. According to the data displayed in Table 1, the adsorption tended to reach saturation within 30 min, the period of time later used for multilayer assembly. According to the adsorption isotherms of SFC and HTACC shown in Fig. 4, the concentrations of 2 and 3 mg/mL seemed to be sufficient for SFC and HTACC to reach their adsorption equilibrium, respectively. The effect of ionic strength on the polyelectrolyte adsorption was studied using QCM measurement in conventional mode. Frequency changes due to adsorption were measured in air after immersing the gold-coated QCM plate in polyelectrolyte solution for 30 min, and washing and drying. The frequency shifts shown in Table 2 indicate that the amount of polyelectrolyte adsorption increased when NaCl was added. This result can be explained by the fact that the polyelectrolyte tends to adopt a coil-like conformation in the presence of NaCl due to a relaxation of intra-chain and inter-chain electrostatic repulsion in a polymer solution. The adsorbed layer of coil-like polyelectrolyte was therefore thicker than that of the polyelectrolyte favorably exists in the extended form in the absence of NaCl. 3.3. Polyelectrolyte multilayer assembly The conventional mode of QCM measurement was used to follow the multilayer formation on gold-coated QCM plates by monitoring the frequency change as a function of the number of depositions. Three polycation–polyanion pairs of polyelectrolytes were used for multilayer assembly: CHI–PSS, PAH– SFC, and HTACC–PAA. Their structures are illustrated in Fig. 5. It should be noted that all multilayer assembly was carried out in the presence of 1 M NaCl. Progressive increases

336

S. Channasanon et al. / Journal of Colloid and Interface Science 316 (2007) 331–343

Table 1 Frequency shift (Hz) and adsorption time at equilibrium (min) of SFC and HTACC obtained from in situ monitoring by QCM measurement Charged derivative

Concentration (mg/mL)

Frequency shift (Hz)

Adsorption time at equilibrium (min)

SFC

1.0 2.0 3.0

56 747 1580

7 23 4

1.0 2.0 3.0

169 453 1219

29 29 17

HTACC

the quartz (5 × 106 Hz), A is the electrode area (0.3165 cm2 ), ρq is the density of the quartz (2.65 g/cm3 ), and μq is the shear modulus of the quartz (2.95 × 106 N/cm2 ). Substituting these values into Eq. (1) for one-sided adsorption on the QCM plate, we obtain the following relationship between adsorbed mass and frequency shift: m(g) = −5.60 × 10−9 F (Hz).

Assuming a rigid adsorbed layer, homogeneous surface coverage and a polyion layer density (ρ) of 1.2 g/cm3 [28], the thickness (d) of the film (adsorbed on one side of the QCM plate, A = 0.3165 cm2 ) can be estimated by substituting ρdA for m in Eq. (2). The relationship between the thickness and the frequency shift can be expressed as follows: d (nm) = −0.147F (Hz).

Fig. 4. Adsorption isotherms of SFC (") and HTACC (!) determined by QCM using conventional mode. Table 2 Frequency shift (Hz) after adsorption of SFC and HTACC measured by QCM using conventional mode Concentration of charged derivatives

NaCl (M)

Frequency shift (Hz)

2 mg/mL of SFC

0 1.0 0 1.0

28 127 32 145

3 mg/mL of HTACC

in QCM frequency shift with the number of deposition steps (shown in Figs. 6a–6b) evidently indicated a stepwise deposition in all three systems. The QCM measurements in air using the conventional mode allow the dried mass of the deposited polyelectrolyte, which is correlated with the thickness, to be directly calculated from the Sauerbrey equation without having to take into account the swelling, frictional factor or over-adsorption that are usually encountered in the in situ mode. QCM is an extremely sensitive mass sensor, capable of measuring subnanogram levels. The piezoelectric quartz crystal changes its fundamental oscillation frequency (Fq ) as mass is deposited onto (or depleted from) the crystal surface in accordance with the Sauerbrey equation [27]: F =

−2Fq2 m A(ρq μq )1/2

,

(2)

(1)

where F is the change in resonant frequency due to an adsorbed mass, m, Fq is the fundamental resonant frequency of

(3)

It has been demonstrated that a “clean” gold surface is hydrophilic but becomes rapidly hydrophobic because of “contamination” after exposure to the laboratory air even for a short period of time. Hydrophobicity cannot be avoided even after treatment with piranha solution [29]. Thus in our system, the major driving force for the adsorption of the first layer (HTACC, CHI or PAH) on the clean gold-coated QCM plate should be hydrophobic, another interaction that is known to be involved in multilayer assembly [30]. Such a nonspecific interaction, however, yields a slightly thicker first layer and subsequent multilayer structure than ionic interactions (COO− + vs NH+ 3 /N(CH3 )3 ) and/or H-bonding (OH vs NH2 ) between the thiol-modified gold-coated QCM plate carrying carboxyl (COOH) and hydroxyl (OH) groups and the first layer of adsorbed polyelectrolyte (Fig. 6b). This may be attributed to the fact that the polyelectrolyte chains, in the latter case, prefer to adsorb as trains over loops, tails or coil conformations, particularly, when the major driving forces are ionic interactions. The opposite behavior can be expected for the hydrophobic interactions. Despite the lower thickness, the multilayer films assembled on the thiol-modified gold-coated QCM plates should be more stable than those on the gold-coated QCM plates. The average frequency shifts and the corresponding bilayer thicknesses calculated from Eq. (3) are outlined in Table 3. It should be stressed that a thiol-modified gold-coated QCM plate containing a mixed composition of COOH and OH groups with a water contact angle of 55◦ , represents a model of treated PET substrates, the substrates later used for multilayer assembly and protein adsorption studies. The fact that the water contact angle closely resembles that of the treated PET substrate after the pre-treatment (54◦ ) suggests that the thiol-modified gold-coated QCM plate can well mimic the treated PET substrate. Two average frequency shift values were reported in Table 3. Fodd is an average value calculated from the shift between the odd layers (5, 7, 9), and Feven is an average value calculated from the shift between the even layers (6, 8, 10). According to AFM images, shown later in Fig. 8, it can be assumed that the substrates was entirely covered by the multilayer films when the number of layers are beyond 5. No patches or aggregates were observed on any substrates. This is the reason why the thickness estimation did not take the initial 5 layers into consideration.

S. Channasanon et al. / Journal of Colloid and Interface Science 316 (2007) 331–343

337

Fig. 5. Polycations and polyanions used for multilayer assembly.

(a)

(b)

Fig. 6. Frequency shift (Hz) due to multilayer assembly on (a) gold-coated QCM plate and (b) thiol-modified gold-coated QCM plate, obtained from QCM analysis as a function of the number of depositions.

From careful observations of the frequency shift, it is obvious that the shifts between the even layers are always greater than those between the odd layers for the PAH–SFC and CHI– PSS systems. This implies that there may be some changes of the internal structures of the even layers (SFC or PSS) due to relaxation or reconformation toward the interface upon the deposition of the odd layer (PAH or CHI), causing the multilayer to collapse and become more rigid. These changes were found to be highly dependent on the polyelectrolyte present in the outermost layer [31]. The verification of such speculation demands detailed investigation on viscoelastic properties and

internal structures of the multilayer film using a quartz crystal microbalance with dissipation (QCM-D), which is beyond the scope of this research. As shown in Table 3, the frequency change as well as the thickness of the multilayer film built up from HTACC and PAA was the lowest. This can be explained as a consequence of HTACC as a strong polyelectrolyte being adsorbed first as a thin layer. At pH 7, the second layer adsorption was mainly driven by an electrostatic attraction between N(CH3 )+ 3 groups of HTACC and COO− groups of PAA. Although PAA is a weak polyelectrolyte, its adsorption depends strongly on the thick-

338

S. Channasanon et al. / Journal of Colloid and Interface Science 316 (2007) 331–343

Table 3 Average frequency shift and thickness of deposited multilayer film analyzed by QCM measurements using conventional mode Substrate

Multilayer system

Average Fodd (Hz)

Bilayer thickness, db,odd (nm)

Average Feven (Hz)

Bilayer thickness, db,even (nm)

Gold-coated QCM plate

HTACC–PAA CHI–PSS PAH–SFC

21 81 215

3.1 11.9 31.6

16 87 266

2.4 12.8 39.1

Thiolmodified gold-coated QCM plate

HTACC–PAA CHI–PSS PAH–SFC

14 60 191

2.1 8.8 28.1

14 72 235

2.1 10.6 34.5

ness of the first deposited layer of HTACC at neutral pH where COOH groups are essentially ionized to COO− groups. At pH 4, most of the amino groups (NH2 ) of CHI, which is a weak polyelectrolyte, were converted to NH+ 3 . The adsorption of PSS, a strong polyelectrolyte, on CHI was mostly driven by − electrostatic attraction between NH+ 3 groups and SO3 groups. The situation for the CHI–PSS pair is quite similar to that for the HTACC–PAA pair due to two facts: (1) the stepwise adsorption occurs between a pair of strong and weak polyelectrolytes, (2) the adsorption is mainly driven by electrostatic interaction. For that reason, the overall thickness of the multilayer assembled from CHI and PSS was not much greater than that of the multilayer assembled from HTACC and PAA. Unlike the other two polyelectrolyte pairs, the increment of frequency change due to PAH and SFC deposition was unexpectedly high, suggesting that a relatively thick multilayer film was formed. As a result of the incomplete substitution of N -sulfofurfuryl groups at the NH2 position of chitosan (∼60%), SFC bears both SO− 3 and NH2 groups. At pH 8, the adsorption of SFC on the PAH layer and vice versa was therefore driven by both electrostatic − interaction (NH+ 3 vs SO3 ) and H-bonding (NH2 vs OH/NH2 ). Besides the train conformation, the polyelectrolytes were also in the forms of loops, tails, or coils making the adsorbed layer thicker. The unique combination of PAH and SFC then gave rise to the presumably thicker multilayer assembly. Poly(ethylene terephthalate) (PET) has been recognized as a potential polymeric material for biomedical and biomaterials applications. However, due to the lack of bioactive functionality, a number of surface modification methods have been introduced to improve its favorable response to cells or proteins for specific applications. In this study, plasma-treated PET plates, which are commercially available, were used as substrates for multilayer assembly. The treated PET substrate was pre-treated by basic hydrolysis prior to use in order to remove some organic dirt that might exist on the substrate and to enhance the hydrophilicity. After the pre-treatment, the water contact angle of the treated PET surface dropped from 67◦ to 53◦ . It should be noted that the virgin, unmodified PET surface usually possesses a water contact angle in the range of 77◦ –80◦ . The basic hydrolysis has been proven to introduce carboxyl (COOH) and hydroxyl (OH) groups to the surface of PET [32]. We attempted to determine the presence of COOH groups on the surface of the treated PET substrates after the pre-treatment by ATR-FTIR analysis. However, this was unsuccessful possibly due to the modified layer being significantly thinner than the sampling

depth of ATR-FTIR analysis (∼1 µm). Nonetheless, we have verified the presence of COOH groups by using the toluidine blue O assay, a commonly known method for the determination of carboxyl group density. Using the method described by Liu et al. [33], the estimated carboxyl group density on the surface of the treated PET substrate after the pre-treatment is 1.46 × 10−8 mol/cm2 . For the treated PET substrate believed to bear COOH and OH groups, which should show a negatively charged surface at appropriate pH values, the first adsorption cycle was carried out in polycation solution (HTACC, CHI or PAH). The treated PET substrate should be able to adsorb polycations via ionic + interactions between COO− and NH+ 3 /N(CH3 )3 or H-bonding between OH and NH2 . If the total number of layers is odd, the last layer adsorbed contains the polycation. If the total number of layers is even, the last layer adsorbed contains the polyanion. The coverage of the multilayer film on the treated PET substrate was verified by ATR-FTIR analysis. Fig. 7 shows ATR-FTIR spectra of treated PET substrates with multilayer films (21 layers). The presence of absorption peaks at ca. 1192 and 1038 cm−1 assigned to the O=S–O stretching and S=O stretching in Figs. 7c–7d signifies the presence of SFC and PSS in the PAH–SFC and CHI–PSS multilayer films, respectively. The signals due to carbonyl stretching (amide I) and N–H bending (amide II) of the glucosamine unit verifies the existence of CHI in the CHI–PSS multilayer film while the signal from N–H bending alone is certainly enough to demonstrate the presence of PAH in the PAH–SFC multilayer film. In contrast, the ATR-FTIR spectrum of the HTACC–PAA multilayer film could almost be superimposed on that of the treated PET substrate, implying that the multilayer deposition was several times thinner than the sampling depth of the ATR-FTIR analysis. This outcome is quite reasonable considering that the thickness of 21 layers of the HTACC–PAA multilayer film should be ∼20 nm, based on the QCM data. The fact that the signal of C=O stretching from the treated PET substrate markedly diminishes after the deposition of the PAH–SFC multilayer film indicates that the deposited film is quite thick. According to QCM analysis, the thickness of 21 layers of PAH–SFC can be as high as ∼360 nm, which is more than one third of the detection limit of the ATR-FTIR technique. Due to the significantly lower thickness of the CHI–PSS multilayer film (∼110 nm) in comparison with the PAH–SFC multilayer film, the signals from the treated PET substrate dominated those from the deposited CHI–PSS multilayer film. Apparently, the broad peak over the region of

S. Channasanon et al. / Journal of Colloid and Interface Science 316 (2007) 331–343

339

Fig. 7. ATR-FTIR spectra of treated PET substrates (a) after pre-treatment and with multilayer films of (b) (HTACC–PAA)10 –HTACC, (c) (CHI–PSS)10 –CHI, and (d) (PAH–SFC)10 –PAH.

O–H and N–H stretching in the range of 3750–2500 cm−1 can be used as an indication of hydrogen bonding within the multilayer films. The coverage of the deposited multilayer films can be revealed from topographic images obtained by AFM analysis. As shown in Fig. 8, the treated PET substrates became increasing rougher after the pre-treatment by basic hydrolysis. The average roughness (rms) increased from 1.761 to 8.863 nm. In the cases of the PAH–SFC and CHI–PSS multilayer systems, the surfaces became smoother after the deposition of only 6– 7 layers. AFM images of the multilayer films having 6 layers [(HTACC–PAA)3 , (CHI–PSS)3 , and (PAH–SFC)3 ] and 7 layers [(HTACC–PAA)3 –HTACC, (CHI–PSS)3 –CHI, and (PAH– SFC)3 –PAH] are illustrated in the top row and bottom row of Fig. 8, respectively. As expected, the rms was almost unchanged after the deposition of the same number of layers in the case of the HTACC–PAA multilayer film. Only a slight decrease in roughness (rms = 7.904) was observed after 21 layers were deposited (AFM images not shown). This evidence implies complete coverage of the multilayer films, considering that the average roughness is larger than the molecular size of the polyelectrolyte and smaller than the average thickness of a single bilayer, except in the case of the HTACC–PAA multilayer films. If the concept of alternate adsorption is valid, the surface properties of the multilayer film should alternately change. In other words, the multilayer film should be stratified. Water contact angle data shown in Fig. 9 confirmed that the assembled film was stratified. The number appearing on the horizontal scale represents the number of depositions. If the number of

layers is odd, the charge of the outermost layer is positive. If the number of layers is even, the charge of the outermost layer is negative. According to the calculation based on QCM analysis, each individual layer is apparently thicker than the sampling depth of the contact angle measurement (a few Å) so that the wettability of the multilayer film should be strongly dictated by the last layer deposited and the influence of the underlying layers should not be observed. Here, we would like to emphasize that the stratification was still observed in the case of the HTACC–PAA multilayer film, despite its extremely thin individual layers. In addition to the protein adsorption data shown in the next section, this can be another indication of the presence of the HTACC–PAA multilayer film on the treated PET substrate, which cannot be demonstrated by ATR-FTIR and AFM analysis. The assembled films having an odd number of deposited layers (positively charged surface) were clearly more hydrophobic than those having an even number of deposited layers (negatively charged surface). 3.4. Protein adsorption Protein adsorption on a material’s surface is generally regarded as a primary event that occurs when the material comes into contact with its biological surroundings and to certain extent can reflect the performance of the material when it is used as a biomedical device or biomaterial. For this reason, the alternate bioactivity of the assembled thin film was then tested against proteins having a varied size and charge under the experimental conditions (pH 7.4). Their physical properties are outlined in Table 4. Although the protein adsorption on a bio-

340

S. Channasanon et al. / Journal of Colloid and Interface Science 316 (2007) 331–343

Fig. 8. AFM images (5 × 5 µm2 ) of treated PET substrates before and after pre-treatment and treated PET substrates with multilayer films of (HTACC–PAA)3 , (CHI–PSS)3 , (PAH–SFC)3 , (HTACC–PAA)3 –HTACC, (CHI–PSS)3 –CHI, and (PAH–SFC)3 –PAH.

Table 4 Physical properties of proteins used for adsorption studies

Fig. 9. Water contact angle of multilayer films assembled on treated PET substrates.

material’s surface is a competitive phenomenon in the actual physiological environment that simultaneously involves multiprotein systems, understanding the response of a single protein is necessary to gain basic knowledge about the correlation between the physical properties of the protein and its adsorption

Physical property

Albumin (BSA)

Fibrinogen (FIB)

γ -Globulin (GLB)

Lysozyme (LYZ)

Mass (Da) Isoelectric point (pH units)

69,000 4.8

340,000 5.5

165,000 6.5

14,600 11.1

behavior. The information should be fundamentally important for comprehending the more complex systems in which many proteins exist. Since the analysis of the adsorbed protein on the multilayer films deposited on the treated PET substrates relies on the desorption by sodium dodecyl sulfate (SDS), a concern was raised as to whether the deposited polyelectrolytes possibly removed by SDS along with the protein, interfering with the analysis. To address this issue, a set of controlled experiments was performed by soaking the treated PET substrate with a deposited multilayer film before protein adsorption in SDS solution. It was found that the absorbance of the SDS solution was essentially zero (lower than or slightly equal to the blank solution), implying that there was no interference with the protein analysis by the desorbed polyelectrolytes.

S. Channasanon et al. / Journal of Colloid and Interface Science 316 (2007) 331–343

341

Fig. 10. Amount of adsorbed BSA per surface area (µg/cm2 ) of (CHI–PSS)n , (PAH–SFC)n , and (HTACC–PAA)n assembled on treated PET substrates.

Fig. 12. Amount of adsorbed GLB per surface area (µg/cm2 ) of (CHI–PSS)n , (PAH–SFC)n , and (HTACC–PAA)n assembled on treated PET substrates.

Fig. 11. Amount of adsorbed FIB per surface area (µg/cm2 ) of (CHI–PSS)n , (PAH–SFC)n , and (HTACC–PAA)n assembled on treated PET substrates.

Fig. 13. Amount of adsorbed LYZ per surface area (µg/cm2 ) of (CHI–PSS)n , (PAH–SFC)n , and (HTACC–PAA)n assembled on treated PET substrates.

Adsorption of BSA on the multilayer films is shown in Fig. 10. Albumin protein is a carboxylic acid-rich protein. Its carboxylic acid group is converted to a negatively charged carboxylate ion at pH 7.4. As expected, the amount of adsorbed BSA on multilayer assemblies having an odd number of layers (7 and 9) and a positive charge was higher than those having an even number of layers (6 and 10) and a negative charge. Such a trend can be explained by the fact that the adsorption was promoted by electrostatic attraction between the positively charged surface of the outermost layer and BSA. On the other hand, the adsorption of the negatively charged BSA was suppressed on the negatively charged surfaces having PSS, SFC or PAA as the top layer due to the electrostatic repulsion. Like BSA, FIB exhibited adsorption behavior in an alternate fashion corresponding to the surface charge of the multilayer film (Fig. 11). The significantly higher quantity of adsorbed BSA in comparison with the adsorbed FIB is due in large part to the variation in the size of protein. Not only can BSA, as a small protein, adsorb on the top surface, but it may also absorb inside the top layer. This may be the reason why the odd-even effect of BSA adsorption was not so strong in the cases of CHI–PSS and

HTACC–PAA, whose individual layers are quite thin according to QCM analysis. It is most likely that the adjacent underlying layer having a charge opposite to that of the top layer plays a role in adsorption. The closer the top layer is to the adjacent underlying layer, the lower the charge density. Under such circumstances, the influence of the charge characteristics of the top layer on protein adsorption becomes weak. Because the assumed permeation into the top layer of the assembled film was somewhat limited for the large protein like FIB, the odd-even effect was quite independent of the system. With a pI value of 6.5, GLB should be negative in charge at pH 7.4. Its adsorption behavior might be expected to follow the trends previously observed for BSA and FIB. The data shown in Fig. 12 indicate that this is only the case for the PAH– SFC system. Having a pI very close to 7.4, GLB may not be strongly charged. As a result, electrostatic interaction may not be a major driving force for adsorption. It is thus believed that the weak charge and the moderate size of GLB are responsible for the absence of an odd-even effect in the systems of CHI– PSS and HTACC–PAA, whose individual layers are quite thin. Since lysozyme (LYZ) is quite a small protein having the low-

342

S. Channasanon et al. / Journal of Colloid and Interface Science 316 (2007) 331–343

est mass among all studied proteins, all three multilayer systems are capable of adsorbing large quantity of LYZ (Fig. 13). The amount of adsorbed LYZ per surface area reached as high as ∼15 µg/cm2 . In comparison with BSA and FIB, LYZ should exhibit adsorption in a reverse manner due to its positively charged character at pH 7.4. The expected alternate adsorption was only evidenced in the PAH–SFC and HTACC–PAA systems. It is suspected that the enzymatic activity of LYZ towards chitosan perhaps accounts for the unpredicted adsorption trend in the CHI–PSS system. The selective protein adsorption behavior of all three multilayer systems against BSA, FIB and GLB, which is mainly governed by electrostatic attraction/repulsion, is in good agreement with the data previously reported by others [34,35]. The adsorbed proteins appear to interact strongly with the outermost polyelectrolyte layer or with the polyelectrolyte/protein layer complex whatever the sign of the charge of the film. This also implies that the large coverage observed even when proteins adsorb on polyelectrolyte films of same surface charge should be due to protein diffusion along the film. The similar behavior was also observed by Ladam and co-workers [36]. In particular, they have found that when the charges of the protein and the multilayer film are opposite, the adsorbed amounts are much larger than the ones observed when proteins and polyelectrolyte films are similarly charged, and the formation of thick protein layers extending up to several times the largest dimension of the protein can be observed. They have proposed that the formation of thick protein layers can be due to cooperative attractive interactions consecutive to a protein ordering on the surface of a polyelectrolyte multilayer. This could then lead to a first protein layer adsorbed in such a way onto an oppositely charged multilayer film as to exhibit itself a net surface charge which again acts as an attractive surface for further proteins from the solution. Subsequent protein layers can thus be similarly attached. However, as the process evolves, the protein layer becomes less structured so that the surface excess charge decreases and the process stops. In our cases, it was also found that the smaller the protein, the greater the amount of adsorption. The amount of adsorbed protein is ranged from high to low in the following order: LYZ > BSA > GLB > FIB. This sizedependent adsorption is another supportive evidence of protein diffusion. The fact that proteins still adsorb onto similarly charged multilayer films is not unexpected because proteins bear on their surfaces domains with both positive and negative surface excess charge. And electrostatic interactions are not the only driving force for protein adsorption. Moreover, it is not excluded that on negatively charged multilayer film, some cationic polyelectrolyte chains can emerge at the outer surface and are thus also able to interact with the proteins. The same is also true for positively charged multilayer film and anionic polyelectrolyte chains. This interpenetrating effect can, in fact, be realized from the odd-even trend in the protein adsorption. Apparently, the thinner the individual layer, the smaller odd-even trend.

4. Conclusions Charged derivatives of chitosan, N -sulfofurfuryl chitosan (SFC) and N -[(2-hydroxyl-3-trimethylammonium)propyl]chitosan chloride (HTACC) are capable of forming multilayer films in a fashion similar to chitosan. According to QCM analysis, the average thickness of the bilayer is arranged in the following order: PAH–SFC > CHI–PSS > HTACC–PAA. The multilayer films assembled on treated PET substrates were found to be stratified. As characterized by AFM, the multilayer film can completely cover the substrate. Alternate bioactivity of the deposited multilayer film was realized from the results of protein adsorption studies. It has been demonstrated that the proteins adsorbed onto the assembled film in a multilayer fashion, implying that the diffusion of the proteins within the multilayer structure has occurred. The fact that both HTACC and SFC are soluble over a broader pH range than and possess different bioactivity from chitosan suggests that these two charged derivatives of chitosan can be potential candidates for biomedical-related applications. Acknowledgments This research is financially supported by a Research Team Promotion Grant from the Thailand Research Fund (RTA4780004). The authors are indebted to Associate Professor Sanong Ekgasit of the Sensor Research Unit, Department of Chemistry, Faculty of Science, Chulalongkorn University for use of ATR-FTIR facility. The appreciation is extended to Stephan T. Dubas of Metallurgy and Materials Science Research Institute, Chulalongkorn University for helpful comments and suggestions. The contact angle goniometer provided by the National Metal and Materials Technology Center (MTEC) is gratefully acknowledged. References [1] G. Decher, Science 277 (1997) 1232. [2] T. Serizawa, M. Yamaguchi, A. Kishida, M. Akashi, J. Biomed. Mater. Res. 67A (2003) 1060. [3] T. Serizawa, M. Yamaguchi, T. Matsuyama, M. Akashi, Biomacromolecules 1 (2000) 306. [4] T. Serizawa, M. Yamaguchi, M. Akashi, Biomacromolecules 3 (2002) 724. [5] J. Fu, J. Ji, W. Yuan, J. Shen, Biomaterials 33 (2005) 6684. [6] B. Thierry, F.M. Winnik, Y. Merhi, M. Tabrizian, J. Am. Chem. Soc. 125 (2003) 7494. [7] L. Richert, P. Lavalle, E. Payan, X.Z. Shu, G.D. Prestwich, J.F. Stoltz, P. Schaaf, J.C. Voegel, C. Picart, Langmuir 20 (2004) 448. [8] Y. Zhu, C. Gao, T. He, X. Liu, J. Shen, Biomacromolecules 4 (2003) 446. [9] Y. Liu, T. He, C. Gao, Colloids Surf. B 46 (2005) 117. [10] K. Cai, A. Rechtenbach, J. Hao, J. Bossert, K.D. Jandt, Biomaterials 26 (2005) 5960. [11] Q. Chen, J. Han, H.B. Shi, B.Y. Wu, X.H. Xu, T. Osa, Sensors Lett. 2 (2004) 102. [12] M. Yang, Y. Yang, H. Yang, G. Shen, R. Yu, Biomaterials 27 (2006) 246. [13] J. Benesch, P. Tengvall, Biomaterials 23 (2002) 2561. [14] Y. Okamoto, R. Yano, K. Miyatake, I. Tomohiro, Y. Shigemasa, S. Minami, Carbohydr. Polym. 53 (2003) 337. [15] M.M. Amiji, Colloids Surf. B 10 (1998) 263. [16] P. Vongchan, W. Sajomsang, D. Subyen, P. Kongtawelert, Carbohydr. Res. 337 (2002) 1239. [17] Z. Jia, D. Shen, W. Xu, Carbohydr. Res. 333 (2001) 1.

S. Channasanon et al. / Journal of Colloid and Interface Science 316 (2007) 331–343

[18] C.H. Kim, K.S. Choi, J. Ind. Eng. Chem. 8 (2002) 71. [19] C.H. Kim, J.W. Choi, H.J. Chun, K.S. Choi, Polym. Bull. 38 (1997) 387. [20] E.I. Rabea, M.E.-T. Badawy, C.V. Stevens, G. Smagghe, W. Steurbaut, Biomacromolecules 4 (2003) 1457. [21] L.-Y. Zheng, J.-F. Zhu, Carbohydr. Polym. 54 (2003) 527. [22] Y.H. Kim, H.M. Choi, J.H. Yoon, Textile Res. J. 68 (1998) 428. [23] H.-S. Seong, H.S. Whang, S.-W. Ko, J. Appl. Polym. Sci. 76 (2000) 2009. [24] Y.H. Kim, C.W. Nam, J.W. Choi, J.H. Jang, J. Appl. Polym. Sci. 88 (2003) 1567. [25] S.-H. Lim, S.M. Hudson, Carbohydr. Res. 339 (2004) 313. [26] H. Sashiwa, Y. Shigemasa, Carbohydr. Polym. 39 (1999) 127. [27] A. Baba, F. Kaneko, R.C. Advincula, Colloids Surf. A 173 (2000) 39. [28] G. Sauerbrey, Z. Phys. 155 (1959) 206.

343

[29] J. Brandrup, E. Immergut (Eds.), Polymer Handbook, Wiley, New York, 1975. [30] N.J. Geddes, E.M. Paschinger, D.N. Furlong, F. Caruso, C.L. Hoffman, J.F. Rabolt, Thin Solid Films 260 (1995) 192. [31] E. Lojou, P. Bianco, Langmuir 20 (2004) 748. [32] S.M. Notley, M. Eriksson, L. Wgberg, J. Colloid Interface Sci. 292 (2005) 29. [33] W. Chen, T.J. McCarthy, Macromolecules 30 (1997) 78. [34] Y. Liu, T. He, C. Gao, Colloids Surf. B 46 (2005) 117. [35] M. Müller, T. Rieser, P.L. Dubin, K. Lunkwitz, Macromol. Rapid Commun. 22 (2001) 390. [36] G. Ladam, P. Schaaf, F.J.G. Cuisinier, G. Decher, J.-C. Voegel, Langmuir 17 (2001) 878.