Supramolecular interactions between β-cyclodextrin and hydrophobically end-capped poly(ethylene glycol)s: A quartz crystal microbalance study

Journal of Colloid and Interface Science 315 (2007) 800–804 www.elsevier.com/locate/jcis

Note

Supramolecular interactions between β-cyclodextrin and hydrophobically end-capped poly(ethylene glycol)s: A quartz crystal microbalance study Khémara Kham a , Mohamed Guerrouache a , Benjamin Carbonnier a,∗ , Mathieu Lazerges b , Hubert Perrot b , Marie-Claude Millot a a Institut Chimie et Matériaux Paris Est—Equipe Systèmes Polymères Complexes, CNRS-Université Paris 12 UMR 7182, 2-8 rue Henri Dunant,

94320 Thiais, France b Laboratoire Interfaces et Systèmes Electrochimiques, UPR 15 du CNRS, Université Pierre et Marie Curie, 4 place Jussieu, 75252 Paris cedex 05, France

Received 16 February 2007; accepted 29 June 2007 Available online 19 July 2007

Abstract In this study, the supramolecular interactions occurring between β-cyclodextrin-based surfaces and macromolecular chains modified at one end with naphthyl, adamantyl, or phenyladamantyl hydrophobic groups were investigated by means of a quartz crystal microbalance. β-Cyclodextrin-functionalized gold electrodes were obtained through the amide-coupling reaction between mono-6-deoxy-6-amino-β-cyclodextrin and 11-mercaptoundecanoic acid self-assembled monolayer allowing the reproducible preparation of densely grafted surfaces with host properties. The interaction data obtained for the three different modified poly(ethylene glycol)s are in good agreement with our previous studies performed by high performance liquid chromatography and surface plasmon resonance. This evidences that the driving force for the supramolecular interaction is based on the inclusion of the hydrophobic terminal group of the chains within the cyclodextrin cavities. The reversibility of the inclusion process was proven through the regeneration of the original host properties of the sensing surfaces using sodium dodecylsulfate as a competitor for the desorption of the poly(ethylene glycol) chains. © 2007 Elsevier Inc. All rights reserved. Keywords: Quartz crystal microbalance; β-Cyclodextrin; Poly(ethyleneglycol); Inclusion complexe; Surface regeneration

1. Introduction β-Cyclodextrin (β-CD) is a water soluble cyclic oligosaccharide made of seven glucose units linked through α-1,4 bonds [1]. This cone-shaped molecule, usually referred as βCD cavity, exhibits two polar external faces, while the interior of the cavity, containing two rings of C–H groups with a ring of glycosidic oxygen in between, is relatively nonpolar. On account of this specific geometric shape and chemical anisotropy, β-CD molecules may form inclusion complexes in aqueous media with a wide range of molecules or macromolecules containing hydrophobic moieties [2–4]. It was notably shown that similar host–guest interactions occurred in solution through mixing β-CD containing polymers (poly-β-CD) and poly(ethylene ox* Corresponding author. Fax: +33 (0)1 49 78 12 08.

E-mail address: [email protected] (B. Carbonnier). 0021-9797/$ – see front matter © 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.jcis.2007.06.086

ide)s (PEO)s bearing hydrophobic end-groups. Depending on the PEO topology (linear vs star-like) and the number of hydrophobic end-groups per chain (from 2 to 8), associating systems with varied microscopic and macroscopic properties were obtained. More recently, efforts have been devoted to characterize the nature of the interaction between poly-β-CD coated silica and gold surfaces with adamantyl (Ad)-, phenyladamantyl (PhAd)-, or naphthyl (Nap)-end-capped poly(ethylene glycol)s using high performance liquid chromatography [5] and surface plasmon resonance [6] techniques, respectively. In both cases, it was clearly demonstrated that the driving force for the interfacial interaction relies on formation of inclusion complexes. Herein, stability and reversibility of the association between β-CD molecules immobilized on gold surfaces and hydrophobically modified methoxy poly(ethylene glycol)s (MPEG-R)s are investigated using the quartz crystal microbalance (QCM). The (MPEG-R)s consist of linear chains bearing naphthyl, adamantyl, or phenyladamantyl groups as hydrophobic modi-

K. Kham et al. / Journal of Colloid and Interface Science 315 (2007) 800–804

fiers at only one end. The QCM is often used as a complementary method of surface plasmon resonance (SPR) to monitor in situ interfacial phenomena with a high sensitivity [7,8]. While SPR gives information about the optical thickness, the QCM is an acoustic method for which the shift in the resonance frequency (F ) of the quartz crystal may be considered, in a first approximation, as proportional to any adsorbed mass. 2. Experimental 2.1. Materials Chemically modified (MPEG-R)s (Mw = 5000 g mol−1 ) bearing naphthyl (Nap) and phenyladamantyl (PhAd) as hydrophobic end-groups were obtained according to the previously reported method [9]. The adamantyl-substituted polymer (MPEG-Ad) was obtained through a similar esterification reaction using 1-adamantoyl chloride as a reagent (Sigma–Aldrich, St. Quentin Fallavier, France). 11-mercaptoundecanoic acid (MUA), N -hydroxysuccinimide (NHS), 1-[3-(dimethylamino)propyl]-3-ethylcarbodiimide hydrochloride (EDC) and sodium dodecylsulfate (SDS) were also purchased from Sigma–Aldrich (St. Quentin Fallavier, France) and were used as received. Sulfuric acid (95%) and hydrogen peroxide (30%) were obtained from VWR and Carlo Erba Reagent, respectively. The MUA solution (10−3 mol L−1 ) was prepared in ethanol (99.8%, Fluka) while aqueous solutions of MPEG or MPEG-R (0.1 g L−1 ) and SDS (10 g L−1 ) were used. Water used in all experiments was deionized and double distilled. 2.2. Synthetic methods 2.2.1. Synthesis of mono-6-deoxy-6-amino-β-cyclodextrin (β-CDNH2 ) Introduction of amino group into β-CD was preformed through the sequential synthesis of mono-6-deoxy-6-(p-toluenesulfonyl)-β-cyclodextrin, mono-6-deoxy-6-azido-β-cyclodextrin and mono-6-deoxy-6-amino-β-cyclodextrin [10–12]. Purity of the β-CDNH2 derivative was checked by high performance liquid chromatography. 2.2.2. Immobilization of β-CDNH2 onto gold electrodes Just prior to use, the QCM surface was cleaned with piranha solution (a 3:1 mixture of H2 SO4 :H2 O2 ) followed by a rinsing step with water. In this work, a three-step procedure was used to prepare the host monolayers: (i) COOH-functionalized gold surface was prepared through the adsorption of the thiol end of the bifunctional MUA reagent [6]. Thus the gold electrode was rinsed with ethanol and dried under argon atmosphere; (ii) activation of the surface was performed by converting the carboxylic groups into NHS-ester by reaction with NHS (0.05 mol L−1 ) and EDC (0.2 mol L−1 ) during ∼20 min [13]; (iii) finally, the activated surface was exposed to an aqueous solution of β-CDNH2 (10 g L−1 ) resulting in the covalent binding of β-CD cavities through amide link. The steps (ii) and (iii) were conducted within the QCM cell and the frequency shift (F ) was calculated after rinsing step to monitor in situ the

801

amide-forming coupling reaction. A schematic representation of the different synthetic steps is shown in Figs. 1a and 1b. 2.2.3. Interfacial interaction between (MPEG-R)s and immobilized β-CDNH2 Interfacial events between β-CD cavities and hydrophobic chain-ends (Fig. 1c) were investigated by measuring the changes in frequency as a function of time. Both adsorption and desorption processes were monitored in real time while flowing aqueous MPEG-R (0.1 g L−1 ) and SDS (10 g L−1 ) solutions, respectively. The specificity of the interaction between the MPEG-R polymers and the CD cavities was proved by injecting an aqueous solution of unmodified MPEG (0.1 g L−1 ). 2.3. QCM apparatus The efficiency of the specific interaction between β-CD and modified (MPEG-R)s was investigated by gravimetric measurement by QCM. In fact, the frequency responses of QCM devices in aqueous solutions may be correlated to interfacial reactions involving changes in mass. The QCM is an oscillating quartz crystal whose frequency response to a mass change was described by the Sauerbrey equation [14] for thin, uniform and purely added layers as follows:   m , F = −2F02 (1) nA(μρ)1/2 where F0 is the fundamental quartz frequency (9 MHz), F is the measured shift in frequency (Hz), A is the piezoelectrically active area of the crystal (cm2 ), ρ is the quartz density (2.648 g cm−3 ), μ is the shear wave velocity of the AT-cut quartz used (2.947 × 1011 g s−2 cm−1 ), m is the mass change (g) on the surface of the crystal, n is the overtone number (n = 3). AT-cut planar quartz crystals of 14 mm diameter with a 9 MHz nominal resonance frequency (Matel-Fordhal, France) were used. Two identical gold electrodes, 2000 Å thick and 5 mm in diameter were deposited by evaporation techniques on both sides of crystals with a 250 Å chromium underlayer. The resonators were connected with a silver conducting paste, through wires, to a BNC adaptator. A home-made oscillator was designed to drive the crystal at 27 MHz which corresponds to the third overtone of the 9 MHz quartz resonator. To improve the stability, all the electronic oscillator components were temperature-controlled by a Watlow heater current monitor with stability better than 0.1 ◦ C (30 ± 0.1 ◦ C). An experimental cell was developed: the crystal was mounted between two O-ring seals inserted in a plexiglass cell. Only one face of the quartz was in contact with the solutions. The cell volume was 50 µl. The apparatus included a P1 micropump (Pharmacia) to assure a 50 µl min−1 constant flow of the solutions. The frequency was computer-controlled by home-made software in C language and measured with a Fluke PM 6685 frequency counter. Concerning the validity of the Sauerbrey equation, viscoelastic effects were neglected in our experiments as the thickness of the polymer was considered acoustically thin enough [15] and surface concentration Γ (mol cm−2 ) can be calculated as follows:

802

K. Kham et al. / Journal of Colloid and Interface Science 315 (2007) 800–804

Fig. 1. Schematic representation showing the overall layer construction of the β-CDNH2 functionalized surface and the subsequent interaction with MPEG-R through inclusion process.

m n (2) = , MA A where n and M are the mole number and molar mass of the molecules adsorbed on the surface, respectively. From the surface concentration values calculated for the cyclodextrin and MPEG-R layers, the degree of complexation (τ ) of β-CDNH2 cavities can be estimated as described in Eq. (3). This gives an estimate of the amount of β-CD cavities interacting with hydrophobically modified poly(methoxyethylene glycol)s: Γ =

τ=

nMPEG-R FMPEG-R Mβ -CD = . nβ -CD Fβ -CD MMPEG-R

Table 1 Variation of frequency (F ) measured in situ for the grafting of β-CDNH2 and the subsequent interactions with the three hydrophobically end-capped poly(methoxyethylene glycol)s. Calculated values of surface concentration (Γ , Eq. (2)) and β-CDNH2 complexation degree (τ , Eq. (3)) Γ (10−11 mol cm−2 )

R

F β-CDNH2

MPEG-R

β-CDNH2

MPEG-R

PhAd Ad Nap

76 ± 7 76 ± 7 76 ± 7

128 ± 6 74 ± 7 14 ± 3

11 ± 1 11 ± 1 11 ± 1

4.3 ± 0.3 2.4 ± 0.3 0.4 ± 0.1

τ 0.39 0.22 0.04

(3)

3. Results and discussion 3.1. Immobilization of β-CDNH2 onto gold electrodes CD-based monolayers may be prepared either by the Langmuir–Blodgett [16] or self-assembling [17] methods. In the latter case, thiol- or sulfur-containing β-CD derivatives have been mainly used and the structural parameters of the organized monolayer were discussed with respect to the chemical structure of the CD molecules [18–20]. The initial parts of the sensorgrams presented in Figs. 2a–2c display the time dependence of the frequency corresponding to the flow of aqueous β-CDNH2 solution over the NHS treated MUA surface, and subsequent rinsing (steps 1 and 2). The coupling reaction leads to a neat decrease of the frequency associated with the increase of the adsorbed mass. The average response recorded for the

three β-CDNH2 modified gold surfaces prepared for the further investigation of the interactions with the MPEG-R bearing hydrophobic (adamantyl, phenyladamantyl, and naphthyl) end groups is 76 Hz (Table 1). This suggests that rather homogeneous and reproducible β-CDNH2 modified surfaces are formed in the equilibrium regime. Based on these results and on the linear relationship between the frequency shift and the adsorbed mass, a surface concentration for β-CDNH2 of about 1.1 × 10−10 mol cm−2 is assumed. Considering monolayers of closely packed β-CD cavities, surface density from 7.5 × 10−11 up to 1.9 × 10−10 mol cm−2 can be obtained for 2D hexagonal lattices with a face-on or an edge-on arrangement of the β-CD molecules, respectively [18,21]. Although it is not possible to draw final conclusions from these simple measurements, it is interesting to note that the synthetic pathway we applied, i.e., grafting of β-CDNH2 molecules onto a functional SAM, leads to rather densely packed surface structure.

K. Kham et al. / Journal of Colloid and Interface Science 315 (2007) 800–804

803

Fig. 3. Dependence of the QCM response (F ) versus time monitored in situ for adsorption/desorption cycles recorded for (a) MPEG-Ad and (b) MPEGPhAd. Step 1: adsorption of MPEG-R (0.1 g L−1 ) onto the β-CD-grafted gold surface. Step 2: rinsing with water. Step 3: rinsing with SDS solution (10 g L−1 ).

Fig. 2. (a–c) Dependences of the QCM response (F ) versus time monitored in situ for the immobilization of β-CDNH2 (10 g L−1 ) on a MUA functionalized surface after NHS/EDC activation and the interaction between the host surface and (a) MPEG-PhAd, (b) MPEG-Ad, and (c) MPEG-Nap. Step 1: flowing of β-CDNH2 solution (10 g L−1 ) onto activated MUA-functionalized gold surface. Step 2: rinsing with water. Step 3: adsorption of MPEG-R (0.1 g L−1 ) onto the CD-grafted gold surface. (d) Control experiment showing the interaction between the host surface and unmodified MPEG. Step 1*: adsorption of MPEG (0.1 g L−1 ) onto the CD-grafted gold surface. Step 2: rinsing with water.

3.2. Interaction between β-CDNH2 modified electrodes and (MPEG-R)s The characterization of the guest properties of the CD surfaces was performed with polymer solutions at a concentration of 0.1 g L−1 . As representative examples, Figs. 2a–2c display the QCM responses recorded for the three types of hydrophobically modified (MPEG-R)s. The flowing of the polymer solution over the biosensor surface induces a decrease of the frequency that levels off after the rinsing step. To exclude any contribution from physical adsorption, i.e., nonspecific interaction between the MPEG and the β-CD-based surface, to the above mentioned frequency decrease, aqueous solution of unmodified poly(methoxyethylene glycol) was flowed over the β-CD surface. In this case, no significant change in the frequency value was observed after rinsing, suggesting that the contribution of the methoxy end group to the formation of inclusion complex may be neglected (Fig. 2d). Similarly, neither modified nor unmodified PEG showed ability to adsorb on bare gold surface. From these results, formation of only 1:1 type inclusion complexes between hydrophobic MPEG-R and CD is assumed. In Fig. 2 it is also seen that the frequency shift, that is the complex stability, is strongly dependent on the nature of the end group (Table 1). It is well-established that para-

meters such as size, geometry and hydrophobic character of the complexed moieties may influence drastically the inclusion process. From a qualitative point of view, the increase in the F values from the naphthyl- through the adamantyl- and to the phenyladamantyl-modified (MPEG) is presumed to reflect an increase in the affinity constants that is in good agreement with the previously reported results [5,6]. 3.3. Reversibility of the interaction between β-CDNH2 modified electrodes and (MPEG-R)s The reversibility of the interaction is a key parameter to allow the design of a wide variety of original and regenerable CD-based systems through interactions with guest macromolecules of varied functionalities. Considering inclusion complexes formation, the interaction can be reversed upon addition of a competitor agent such as an organic modifier or a surfactant [6]. Among these competitor agents, SDS exhibits a strong interaction with β-CD cavities, with a more pronounced effect above the critical micelle concentration [22,23]. Fig. 3 shows the profile of the QCM response versus time during adsorption/desorption cycles performed with MPEG-Ad and MPEGPhAd solutions. The neat variations of the QCM response after the rinsing step are 69 and 127 Hz for MPEG-Ad and MPEGPhAd, respectively, to be compared with the values in Table 1 obtained with newly prepared β-CD surfaces. The good concordance between these frequency variations clearly demonstrates the nearly total regeneration of the QCM surface in the presence of the surfactant molecules at a concentration higher than the critical micellar concentration. 4. Conclusion In this work, the quartz crystal microbalance was successfully applied for the study of the interactions involved between β-cyclodextrin-functionalized electrodes and poly(methoxyethylene glycol) macromolecules terminated by a hydrophobic

804

K. Kham et al. / Journal of Colloid and Interface Science 315 (2007) 800–804

moiety. Effect of various interfacial events such as the covalent coupling of β-CDNH2 , adsorption of the hydrophobic MPEGR as well as the sensing surface regeneration, on the frequency variation of the QCM has been monitored in situ. A major experimental result is that the frequency variation, expressing the stability of the interaction, is strongly function of the nature of the hydrophobic terminal group and increases from the naphthyl-, through the adamantyl- to the phenyladamantylsubstituted MPEG-R. For the three polymers, nearly similar frequency variations were obtained for newly prepared CD surfaces and regenerated ones demonstrating the possibility for the repetitive use of the functionalized surface. References [1] W. Saenger, Angew. 92 (1980) 343. [2] Y. Matsui, T. Nishioka, T. Fujita, Top. Curr. Chem. 128 (1985) 61. [3] C. Amiel, A. Sandler, B. Sébille, P. Valat, V. Wintgens, J. Polym. Anal. Character. 1 (1999) 105. [4] C. Amiel, B. Sébille, Adv. Colloid Interface Sci. 79 (1999) 105. [5] C. David, M.C. Millot, B. Sébille, J. Chromatogr. B 753 (2001) 93. [6] C. David, M.C. Millot, B. Sébille, Y. Lévy, Sens. Actuat. B 90 (2003) 286.

[7] X. Su, Y.-J. Wu, W. Knoll, Biosens. Bioelectr. 21 (2005) 719. [8] X. Su, Y.-J. Wu, R. Robelek, W. Knoll, Langmuir 21 (2005) 348. [9] C. David, M.C. Millot, L. Brachais, C. Amiel, B. Sebille, Macromol. Rapid Commun. 21 (2000) 990. [10] L.D. Melton, K.N. Slessor, Carbohydr. Res. 18 (1971) 29. [11] K. Takahashi, K. Hattori, F. Toda, Tetrahedron Lett. 25 (1984) 3331. [12] R.C. Petter, J.S. Salek, C.T. Sikorsky, G. Kumavarel, F.T. Lin, J. Am. Chem. Soc. 117 (1990) 3860. [13] B. Johnsson, S. Löfas, G. Lindquist, Anal. Biochem. 198 (1991) 268. [14] G. Sauerbrey, Z. Phys. 155 (1959) 206. [15] L. Banda, M. Alcoutlabi, G.B. McKenna, J. Polym. Sci. B 44 (2006) 801. [16] K. Odashima, M. Kotato, M. Sugawara, Y. Umezawa, Anal. Chem. 65 (1993) 927. [17] Y. Maeda, T. Fukuda, H. Yamamoto, H. Kitano, Langmuir 13 (1997) 4187. [18] M.T. Rojas, R. Königer, J.F. Stoddart, A.E. Kaifer, J. Am. Chem. Soc. 117 (1995) 336. [19] G. Nelles, M. Weisser, R. Back, P. Wohlfart, G. Wenz, S. Mittler-Neher, J. Am. Chem. Soc. 118 (1996) 5039. [20] I. Suzuki, K. Murakami, J.I. Anzai, T. Osa, P. He, Y. Fang, Mater. Sci. Eng. C 6 (1998) 19. [21] M. Weisser, G. Nelles, P. Wohlfart, G. Wenz, S. Mittler-Neher, J. Phys. Chem. 100 (1996) 17893. [22] N. Furasaki, H. Yodo, S. Hada, S. Neya, Bull. Chem. Soc. Jpn. 65 (1992) 1323. [23] W.M.Z. Wan Yunus, J. Taylor, D.M. Bloor, D.G. Hall, E. Wyn-Jones, J. Phys. Chem. 96 (1992) 8979.