A Biomimetic Alternative to Poly(ethylene glycol) as an Antifouling

Jul 11, 2008 - Poly(l-lysine)-graft-dextran (PLL-g-dex), graft copolymers with dextran side chains grafted onto a poly(l-lysine) backbone, previously ...
2 downloads 5 Views 699KB Size
Download PDF
8850

Langmuir 2008, 24, 8850-8856

A Biomimetic Alternative to Poly(ethylene glycol) as an Antifouling Coating: Resistance to Nonspecific Protein Adsorption of Poly(L-lysine)-graft-dextran Chiara Perrino,† Seunghwan Lee,† Sung Won Choi,‡ Atsushi Maruyama,‡ and Nicholas D. Spencer*,† Laboratory for Surface Science and Technology, Department of Materials, ETH Zurich, Wolfgang-Pauli-Strasse 10, CH-8093 Zurich, Switzerland, and Institute for Materials Chemistry and Engineering, Kyushu UniVersity, 744-CE11 Motooka, Nishi, Fukuoka 819-0395, Japan ReceiVed March 26, 2008. ReVised Manuscript ReceiVed May 16, 2008 Poly(L-lysine)-graft-dextran (PLL-g-dex), graft copolymers with dextran side chains grafted onto a poly(L-lysine) backbone, previously shown to be effective as stabilizers of DNA triple helices and as carriers of functional genes to target cells or tissues, were employed in this work to prevent nonspecific adsorption of proteins, as determined by means of optical waveguide lightmode spectroscopy. PLL-g-dex copolymers readily adsorb from aqueous solution onto negatively charged oxide surfaces and significantly reduce nonspecific protein adsorption onto bare silica-titania surfaces. While effective and equivalent surface adsorption and antifouling properties were observed for PLL-g-dex copolymers in a variety of architectures, nanotribological analysis by atomic force microscopy was able to distinguish between the different brush densities produced.

Introduction Control of biofouling is the motivation for a large body of ongoing research in many disparate fields. The design of antifouling surfaces for biomedical device applications would allow adverse biological reactions to be avoided, since these generally arise from the conformational transformations undergone by proteins adsorbed on biomaterials. Such reactions can be life threatening and can certainly limit the performance of the devices.1–5 The undesirable adsorption of microorganisms and the formation of biofilms are also detrimental in several nonbiomedical areas: the consequences of biofouling for water purification, transport or storage systems, heat-transfer components, ships’ hulls, and static marine structures are considerable, both economically and ecologically.6–11 Poly(ethylene glycol) (PEG) has been of great interest in the preparation of protein-resistant surfaces, and PEG coatings have proven to successfully reduce or prevent nonspecific protein adsorption, especially when they are in a brush conformation.12–23 PEGsa simple linear polyethersdoes, however, have some limitations: for instance it is susceptible to thermal and oxidative * Corresponding author. E-mail: [email protected]. † ETH Zurich. ‡ Kyushu University. (1) Anderson, J. M.; Cima, M. J.; Langer, R.; Shawgo, R. S.; Shive, M. S.; von Recum, H.; Voskerician, G. Biomaterials 2003, 24(11), 1959–1967. (2) Dai, L.; St John, H. A. W.; Bi, J.; Zientek, P.; Chatelier, R. C.; Griesser, H. J. Surf. Interface Anal. 2000, 29, 46–55. (3) Hook, F.; Vo¨ro¨s, J.; Rodahl, M.; Kurrat, R.; Bo¨ni, P.; Ramsden, J. J.; Textor, M.; Spencer, N. D.; Tengvall, P.; Gold, J.; Kasemo, B. Colloids Surf., B 2002, (24), 155–170. (4) Morra, M. J. Biomater. Sci., Polym. Ed. 2000, 11(6), 547–569. (5) Wang, M. S.; Palmer, L. B.; Schwartz, J. D.; Razatos, A. Langmuir 2004, 20, 7753–7759. (6) Griesser, H. J.; Hartley, P. G.; McHarthur, S. L.; McLean, K. M.; Meagher, L.; Thissen, H. Smart Mater. Struct. 2002, 11, 652–661. (7) Brady, R. F. J. ProtectiVe Coat. Linings 2000, 17(6), 42–46. (8) Rajagopal, S.; Sasikumar, N.; Jakapaul, A.; Nair, K. V. K. Biofouling 1991, 3(4), 311–324. (9) Charmain, J. K.; Osborn, K. S.; Rickard, A. H.; Robson, G. D.; Handley, P. S. Handb. Water Wastewater Microbiol. 2003, 757–775. (10) Flemming, H.-C.; Griebe, T.; Schaule, G. Water Sci. Technol. 1996, 34(5), 517–524. (11) Melo, L. F.; Bott, T. R. Exp. Therm. Fluid Sci. 1997, 14(4), 375–381.

degradation.24–28 The low long-term stability of PEG is one of its major limits and can be responsible for decreased performance in preventing nonspecific protein adsorption. Mimicking biological membrane surfaces provides an alternative approach for conferring antifouling properties to surfaces: the highly hydrated glycocalyx, which surrounds certain kinds of cells, is known, for instance, to possess antiadhesive properties. Carbohydratessthe principal component of the glycocalyxsare thought to be mainly responsible for its ability to prevent undesirable nonspecific adsorption of proteins.29–33 Carbohydrates may therefore constitute a (12) Yu, W. H.; Kang, E. T.; Neoh, K. G. Langmuir 2005, 21(1), 450–456. (13) McGurk, S. L.; Green, R. J.; Sanders, G. H. W.; Davies, M. C.; Roberts, C. J.; Tendler, S. J. B.; Williams, P. M. Langmuir 1999, 15(15), 5136–5140. (14) Lee, J. H.; Kopeckova, P.; Kopecek, J.; Andrade, J. D. Biomaterials 1990, 11, 455–464. (15) Fan, X.; Lin, L.; Messersmith, P. B. Biomacromolecules 2006, 7(8), 2443– 8. (16) Otsuka, H.; Satomi, T.; Itadani, J. H.; Nagasaki, Y.; Okano, T.; Horiike, Y.; Kataoka, K. Eur. Cells Mater. 2003, 6(1), 102. (17) Ma, H.; Hyun, J.; Stiller, P.; Chilkoti, A. AdV. Mater. 2004, 16(4), 338– 341. (18) Heuberger, M.; Drobek, T.; Spencer, N. D. Biophys. J. 2005, 88, 495– 504. (19) Kenausis, G. L.; Vo¨ro¨s, J.; Elbert, D. L.; Huang, N.; Hofer, R.; RuizTaylor, L.; Textor, M.; Hubbell, J.; Spencer, N. D. J. Phys. Chem. B 2000, 104, 3298–3309. (20) Lee, S.; Iten, R.; Mu¨ller, M.; Spencer, N. D. Macromolecules 2004, 37(22), 8349. (21) Pasche, S.; De Paul, S. M.; Vo¨ro¨s, J.; Spencer, N. D.; Textor, M. Langmuir 2003, 19, 9216–9225. (22) Pasche, S.; Vo¨ro¨s, J.; Griesser, H. J.; Spencer, N. D.; Textor, M. J. Phys. Chem. B 2005, 109, 17545–17552. (23) Pasche, S.; Vo¨ro¨s, J.; Spencer, N. D.; Textor, M. Langmuir 2003, 19, 9216. (24) Han, S.; Kim, C.; Kwon, D. Polym. Degrad. Stab. 1995, 47(2), 203–208. (25) Han, S.; Kim, C.; Kwon, D. Polymer 1997, 38(2), 317–323. (26) Madorsky, S. L.; Straus, S. J. Polym. Sci. 1959, 26, 183–194. (27) Reich, L. J. Appl. Polym. Sci. 1969, 13, 977–988. (28) Yang, L.; F., H.; Blease, T. G.; Thompson, R. I. G. Eur. Polym. J. 1996, 32(5), 535–547. (29) Holland, N. B.; Qiu, Y.; Ruegsegger, M.; Marchant, R. E. Nature 1998, 392, 799–801. (30) McArthur, S. L.; McLean, K. M.; Kingshott, P.; St John, H. A. W.; Chatelier, R. C.; Griesser, H. J. Colloids Surf., B 2000, 17(1), 37–48. (31) Piehler, J.; Brecht, A.; Hehl, K.; Gauglitz, G. Colloids Surf., B 1999, 13, 325–336.

10.1021/la800947z CCC: $40.75  2008 American Chemical Society Published on Web 07/11/2008

Nonspecific Adsorption of Proteins

biomimetic alternative to PEG as components of antifouling coatings. Dextran is a natural polysaccharide consisting of an R(1f6)linked glucan with side chains attached to the 3-positions of the backbone glucose units. It is nontoxic, water-soluble, and neutralsfavorable characteristics for application in the biomedical field. Dextran has already been employed for the preparation of protein-resistant surfaces and revealed good antifouling properties.29–31,33–39 Martwiset et al. compared the antifouling properties of dextrans of different molecular weights and different extents of oxidation grafted onto silicon surfaces and reported that the chemical structure of the grafted polymers (relative amount of oxidized groups, molecular weight of dextran) strongly affects their conformation and ability to prevent protein adsorption. They suggested that -OH groups of dextran chains, being both hydrogen bond acceptors and donors, can interact with each other strongly, leading to the collapse of pristine dextran chains, whereas -CHO groups of oxidized dextran only form hydrogen bonds with the surrounding water molecules, thus maintaining ¨ sterberg et al. bound a relatively extended conformation.36 O lightly oxidized dextran in a loops-and-trains configuration on polystyrene and compared its performance in preventing nonspecific protein adsorption to that of dextran/PEG bound in an end-on configuration. They reported that dextran chains grafted with the side-on configuration are much more effective than those with the end-on configuration in reducing protein adsorption.39 However, for the end-on configuration, they did not study the influence of the grafting density on the protein-rejecting ability of the coatings and only a dilute density regime was investigated. In fact it is difficult to achieve high surface brush densities by covalent coupling of polymers onto surfaces (“grafting to”) in an end-on configuration: polymer chains that form well-hydrated random coils in solution will not, in fact, spontaneously adsorb forming dense elongated brushes, due to their hydration shells, which cause mutual steric repulsion. Bosker et al. were the first to consider dextran brushes with high grafting densities: they demonstrated the almost complete protein resistance of densely packed polystyrene-dextran block copolymers (PS-dextran) adsorbed onto hydrophobic surfaces.34 An alternative approach to produce a dense polymer brush involves the grafting of polymer side chains onto a backbone, resulting in the formation of a comblike structure. When adsorbed onto a surface via the backbone, the side chains are forced away from the surface, spontaneously forming a brush, providing the surrounding medium constitutes a good solvent for the side chains. An example of such behavior is seen with poly(L-lysine)-graftPEG (PLL-g-PEG),19,40 a graft copolymer consisting of PEG chains grafted onto a polycationic PLL backbone, which has proven to be highly effective at preventing nonspecific adsorption (32) Qiu, Y.; Zhang, T.; Ruegsegger, M.; Marchant, R. E. Macromolecules 1998, 31(1), 165–171. (33) Sen Gupta, A.; Wang, S.; Link, E.; Anderson, E. H.; Hofmann, C.; Lewandowski, J.; Kottke-Marchant, K.; Marchant, R. E. Biomaterials 2006, 27(16), 3084–3095. (34) Bosker, W. T. E.; Patzsch, K.; Cohen Stuart, M. A.; Norde, W. Soft Matter 2007, 3, 754–762. (35) Frazier, R. A.; Matthijs, G.; Davies, M. C.; Roberts, C. J.; Schacht, E.; Tendler, S. J. B. Biomaterials 2000, 21, 957–966. (36) Martwiset, S. Langmuir 2006, 22, 8192–8196. (37) Mc Lean, K. M.; Johnson, G.; Chatelier, R. C.; Beumer, G. J.; Steele, G.; Griesser, H. J. Colloids Surf., B 2000, 18, 221–234. ¨ sterberg, E.; Bergstro¨m, K.; Holmberg, K.; Riggs, J. A.; Van Alstine, (38) O J. M.; Schuman, T. P.; Burns, N. L.; Harris, J. M. Colloids Surf., A 1993, 77, 159–169. ¨ sterberg, E.; Bergstro¨m, K.; Holmberg, K.; Schuman, T. P.; Riggs, (39) O J. A.; Burns, N. L.; Van Alstine, J. M.; Harris, J. M. J. Biomed. Mater. Res. 1995, 29, 741–747. (40) Huang, N.; Michel, R.; Voros, J.; Textor, M.; Hofer, R.; Rossi, A.; Elbert, D.; Hubbell, J.; Spencer, N. D. Langmuir 2001, 17, 489–498.

Langmuir, Vol. 24, No. 16, 2008 8851

of proteins onto oxide surfaces in an aqueous environment.18,19,21–23,41 PLL-g-PEG copolymers have been shown to readily adsorb from aqueous solutions onto negatively charged surfaces via electrostatic interactions; the positive charges present on the protonated primary amine groups of the PLL backbone in neutral aqueous environment lead to its rapid immobilization onto negatively charged surfaces. The side chains, radially distributed along the backbone in bulk solution in order to minimize steric repulsion, stretch out into the solution once the copolymers have adsorbed on the surface, providing sufficiently close inter-PEG spacing. The copolymers employed in the present study, poly(L-lysine)graft-dextran (PLL-g-dex), have a very similar structure to PLLg-PEG, with the same PLL backbone and dextran substituting for PEG as side chains (Figure 1). Dextran, like PEG, is neutral, and it therefore has no electrostatic interactions with the polycationic backbone. Furthermore it has multiple reactive sites,42 facilitating subsequent functionalization, and is much cheaper than end-functionalized PEG. The undesirable oxidation to peroxides in air, a weakness of PEG-based systems, is not expected to be a problem with sugar chains (the relative longterm stability of dextran and PEG chains as antifouling coatings is currently under investigation). While the PS-dextran brushes studied by Bosker et al. are adsorbed onto hydrophobic substrates via hydrophobic interactions, PLL-g-dex copolymers, similarly to PLL-g-PEG, are expected to spontaneously and electrostatically adsorb onto metal oxide surfaces, thus broadening the range of potential applications of dextran brushes. PLL-g-dex copolymers have previously proven to be effective as stabilizers for DNA triple helices and as potential carriers of functional genes to target cells or tissues.43–47 On the basis of the well-established excellent performance of PLL-g-PEG in preventing nonspecific protein adsorption and the above-mentioned studies of the proteinrepelling capabilities of dextran coatings, the focus of this work is the characterization of both the adsorption and antifouling properties of PLL-g-dex graft-copolymers. Previous nanotribological analysis of PEG brushes have allowed subtle differences in brush structure to be distinguished.48 In this study, this approach has been employed to show differences in dextran brush structure as a consequence of modifications in the PLL-g-dex architecture.

Materials and Methods Poly(L-lysine)-graft-Dextran (PLL-g-dex). Poly(L-lysine)-graftdextran (PLL-g-dex) copolymers were synthesized by a reductive amination reaction of poly(L-lysine)-HBr (13 or 6 kDa, polydispersity 1.3 (for both), Sigma-Aldrich, St. Louis, MO) with dextran (5.9 kDa, polydispersity 1.6, Amersham Bioscience, Uppsala, Sweden). A sodium borate buffer (0.1 M, pH 8.5, 0.4 M NaCl) was used as solvent for the reaction. An approximately 10× molar excess of sodium cyanoborohydride (NaBH3CN) to dextran was used to reduce the unstable Schiff base resulting (41) Pasche, S.; Textor, M.; Meagher, L.; Spencer, N. D.; Griesser, H. J. Langmuir 2005, 21, 6508–6520. (42) Massia, S. P.; Stark, J. J. Biomed. Mater. Res. 2001, 56(3), 390–399. (43) Choi, S. W.; Sato, Y.; Akaike, T.; Maruyama, A. J. Biomater. Sci., Polym. Edn. 2004, 15(9), 1099–1110. (44) Ferdous, A.; Akaike, T.; Maruyama, A. Biomacromolecules 2000, 1, 186–193. (45) Ferdous, A.; Watanabe, H.; Akaike, T.; Maruyama, A. J. Pharm. Sci. 1998, 87(11), 1400–1405. (46) Ferdous, A.; Watanabe, H.; Akaike, T.; Maruyama, A. Nucleic Acid Res. 1998, 26(17), 3949–3954. (47) Maruyama, A.; Watanabe, H.; Ferdous, A.; Katoh, M.; Ishihara, T.; Akaike, T. Bioconjugate Chem. 1998, 9. (48) Yan, X.; Perry, S. S.; Spencer, N. D.; Pasche, S.; De Paul, S. M.; Textor, M.; Lim, M. S. Langmuir 2004, 20, 423–428.

8852 Langmuir, Vol. 24, No. 16, 2008

Figure 1. Structural formula of the PLL-g-dex copolymer. k should be taken as an average value. k + 1 represents the grafting ratio of the polymer.

from the reaction between the terminal dextran aldehyde group and primary amine groups of PLL. The resulting graft copolymers (Figure 1) were isolated by dialysis, in order to free them from unreacted starting materials. The molecular weights of the starting materials, as well as of the graft copolymers, were determined by gel permeation chromatography (JASCO 880 model, Tokyo, Japan) with a multiangle light (GPC-MALS, DawnEOS, Wyatt Technology, Santa Barbara, CA). 1H-NMR spectra were recorded in D2O on a JEOL JNM-EX270 spectrometer (Tokyo, Japan) to determine the composition of the resulting copolymers and the grafting ratio. The notation PLL(x)-g[y]-dex(z) for the copolymers was used to represent the molar mass of PLL in kilodaltons (x) (including the counterions, Br-, as precursor), the molar mass of dextran in kilodaltons (z), and the grafting ratio g[y] (defined as the number of lysine monomers/dextran side chain). Four PLL-g-dex copolymers were employed in this work: while the molecular weight of dextran was kept constant at 5.9 kDa, the grafting ratio (from ca. g[3] to ca. g[10]) and the molecular weight of the PLL backbones (13 and 6 kDa) were varied, yielding four different types of PLL-g-dex (see Table 1). For comparison purposes, a PLL(20)-g[3.0]-PEG(5) copolymer with PEG(5) side chains (molecular weight, 5 kDa) grafted onto a PLL(20) backbone at the grafting ratio of g[3.0], provided by SuSoS AG (Du¨bendorf, Switzerland), has also been employed.19 Optical Waveguide Lightmode Spectroscopy (OWLS). Optical waveguide lightmode spectroscopy (OWLS) was employed to characterize the adsorption properties of the polymers and to evaluate their ability to prevent nonspecific adsorption of proteins. Experiments were performed using an OWLS 110 instrument (Microvacuum, Budapest, Hungary). OWLS is an optical biosensing technique for the in situ label-free analysis of adsorption processes.49 The grating-assisted in-coupling of a He-Ne laser into a planar waveguide allows for a direct online monitoring of macromolecule adsorption. When the in-coupling condition is fulfilled, the light is guided by total internal reflection to the ends of the waveguiding layer, where it is detected by a photodiode detector. The adsorbed mass is calculated from the change in the refractive index in the vicinity of the waveguide surface upon adsorption of molecules.3,49,50 The refractive index increment (dn/dc) of dextran was measured by means of a refractometer and a value of 0.131 was used for all measurements. Since the dn/dc values of dextran and polylysine are very similar, no dn/dc correction was made for the different structures investigated. Prior to the experiments, optical waveguide chips (standard: Si0.75Ti0.25O2 on glass, 1.2 × 0.8 cm2, Microvacuum, Budapest, Hungary) were ultrasonicated in 0.1 M HCl for 10 min, (49) Vo¨ro¨s, J.; Ramsden, J. J.; Csucs, G.; Szendro, I.; De Paul, S. M.; Textor, M.; Spencer, N. D. Biomaterials 2002, 23, 3699–3710. (50) Kurrat, R.; Textor, M.; Ramsden, J. J.; Bo¨ni, P.; Spencer, N. D. ReV. Sci. Instrum. 1997, 68(5), 2172–2176.

Perrino et al. rinsed with ultrapure water (Milli-Q Gradient A10, Millipore SA, Molsheim, France), ultrasonicated in 2-propanol for 10 min, rinsed again with Millipore water, and finally dried under a dry nitrogen stream. The substrates were subsequently cleaned in a UV/ozone cleaner (UV/Clean, model 135500, Boeckel industries Inc., Feasterville, PA) for 30 min. A representative adsorption profile for PLL-g-dex (the case of PLL(6)-g[5]-dex(5.9)) is shown in Figure 2. The cleaned waveguides were placed into the OWLS flow cell and equilibrated by exposing to HEPES buffer solution (10 mM 4-(2-hydroxyethyl)piperazine1-ethanesulfonic acid (Sigma), adjusted to pH 7.4 with 6.0 M NaOH solution) overnight, in order to obtain a stable baseline. The waveguides were then exposed to a polymer solution (0.25 mg mL-1 in a similar HEPES buffer) for at least 30 min, resulting in the formation of a polymer adlayer, and rinsed three times with buffer solution for another 30 min, to exclude the influence of the bulk solution and weakly bound polymer that might contribute to changes in the refractive index at the vicinity of the surface. Low-ionic-strength HEPES was used during the polymeradsorption step, in order to maximize the adsorbed mass. To condition the polymer layer to the ionic strength used in the subsequent serumadsorption step for the evaluation of the protein resistance properties, the optical chips were exposed to a higher-ionic-strength HEPES solution (10 mM HEPES with 150 mM NaCl) before exchanging the buffer back to lower-ionic-strength HEPES. The waveguides were finally exposed to a solution of human serum (Control serum, Precinorm U, Roche, Basel, Switzerland) for 30 min and then rinsed with 10 mM HEPES solution every 15 min for 2 h to check the protein-desorption process. Atomic Force Microscopy (AFM). A conventional beamdeflection-based atomic force microscope was employed to characterize the frictional properties of the PLL-g-dex copolymers adsorbed on SiO2 surfaces on a nanometer scale in an aqueous environment. A commercial SPM scan head (Nanoscope IIIa, MultiMode, Veeco Instruments Inc., Santa Barbara, CA) equipped with a liquid cell/tip holder (Veeco Instruments Inc.) was used, and the fine movement of the sample, placed on top of the piezo scan tube, was controlled by SPM 1000 electronics and SPM 32 software (RHK Technology, Inc., Troy, MI), using an SPM Interface Module (RHK Technology) to interface the scan head and the controller. A commercial silicon nitride AFM tip-cantilever assembly (Veeco) was used as the counterface to the polymermodified SiO2 surfaces. The AFM tip-cantilever assembly was air-plasma cleaned (Plasma Cleaner/Sterilizer, PDC-32G instrument, Harrick, Ossining, NY) for 10 s immediately before the measurements and kept in distilled water prior to use. Silicon wafers (1.2 cm × 1.2 cm) were ultrasonicated in ethanol for 10 min, rinsed with Millipore water, dried under a nitrogen stream, and then oxygen-plasma cleaned for 2 min (Plasma Cleaner/ Sterilizer, PDC-32G instrument, Harrick). After the cleaning procedure, the substrates were incubated in the polymer solution (0.1 mg/mL) for 30 min, rinsed with 10 mM HEPES buffer solution, and then dried under a nitrogen stream. All measurements were performed in 10 mM HEPES buffer solution. The nanotribological properties of PLL-g-dex copolymers were characterized by the acquisition of “friction-vs-load” plots in a number of areas on each sample.51 Briefly, the sample was laterally scanned relative to a fixed tip-cantilever assembly position in a line-scan mode while the sample was simultaneously ramped up and then ramped down in vertical direction to provide a variation in normal load. Both normal and lateral deflection of the tip-cantilever assembly generated from the interaction between the probe tip and the sample surface were detected by a four-quadrant photodiode and interpreted as the normal load (converted based on the manufacturers’ normal spring constant value, kN ) 0.58 N m-1) and the frictional forces (raw photodiode signals), respectively. With the friction forces plotted as a function of normal load, “friction-vs-load” plots were obtained. To ensure a valid comparison of the frictional properties (51) Mu¨ller, M. T.; Yan, X.; Lee, S.; Perry, S. S.; Spencer, N. D. Macromolecules 2005, 38(13), 5706–5713.

Nonspecific Adsorption of Proteins

Langmuir, Vol. 24, No. 16, 2008 8853

Table 1. Synthesized PLL-g-dex Copolymers in This Work (Data Relative to PLL(20)-g[3]-PEG(5) Shown for Comparison Purposes)a polymer

no. of grafted side chains per PLL

no. of free lysines per PLL

percentage of side-chain grafting (%)

mol wt of copolymer (kDa)

PLL(13)-g[3.9]-dex(5.9) PLL(6)-g[5]-dex(5.9) PLL(13)-g[7.1]-dex(5.9) PLL(6)-g[10.2]-dex(5.9) PLL(20)-g[3]-PEG(5)

16 5.7 8.8 2.8 37.9

46.3 23 53.4 25.9 75.7

25.6 20 14.1 9.8 33.3

88 28 78 17 203.9

a The numbers of lysine units for the PLL-HBr (13 kDa), PLL-HBr (6 kDa), and PLL-HBr (20 kDa), before grafting with dextran or PEG, are 62.2, 28.7, and 95.7, respectively. The numbers of monomer units for the PEG (5 kDa) and dex (5.9 kDa) are 113.6 and 36.4, respectively.

Figure 2. Representative adsorption profile for PLL-g-dex copolymer onto a silica-titania waveguide: adsorption of the copolymer (PLL(6)g[5]-dex(5.9)) and subsequent exposure to serum.

of various samples, the same tip-cantilever assembly was used for all the measurements.

Results and Discussion Synthesis and Structural Features. PLL-g-dextran copolymers with a variety of different architectures (different grafting ratios and lengths of the backbone) were successfully synthesized. Characterization by 1H-NMR and GPC-MALS was performed to evaluate the grafting ratios and molecular weights, respectively. The detailed structural features of the synthesized copolymers are shown in Table 1. The molecular weight of dextran chains (5.9 kDa) employed in this work was chosen to be comparable to that of PEG (5 kDa) in the PLL-g-PEG copolymers selected for comparison. The number of monomer units of dex(5.9), 36.4 sugar rings, is, however, significantly lower than that of corresponding EG units of PEG(5), 113.6, resulting in different chain lengths for the different side chains: the fully extended chain length of dextran(5.9), 25.5 nm, is significantly shorter than that of PEG(5), 40.5 nm (based on the molecular length of monomers: 0.7 nm for dextran52 and 0.358 nm for ethylene glycol (EG)53). Adsorption Properties: OWLS. QuantitatiVe Analysis of Polymer Adsorption. Polymer adsorption properties were characterized by means of OWLS. The results of the OWLS experiments (the adsorption profile for PLL(6)-g[5]-dex(5.9) is shown in Figure 2 as example) indicate that the PLL-g-dex copolymers spontaneously adsorb from aqueous solution (10 mM HEPES buffer, pH 7.4) onto metal oxide surfaces. Upon exposure of a waveguide surface to the polymer solution (after 70 min of exposure to HEPES buffer to achieve the baseline), the adsorption process occurred rapidly, such that more than 90% of the final mass of adsorbed polymer was reached within (52) Kawaguchi, T.; Hasegawa, M. J. Mater. Sci.: Mater. Med. 2000, 11, 31–35. (53) Rixman, M. A.; Dean, D.; Ortiz, C. Langmuir 2003, 19, 9357–9372.

Figure 3. OWLS results for the polymer (filled histograms) and serum adsorption (empty histograms) onto bare silica-titania waveguides, including all the investigated PLL-g-dex copolymers (indicated with the length of the backbone and their grafting ratio), PLL(20)-g[3]-PEG(5) and dex.

the first 5 min, and resulted in the formation of a polymer adlayer on the waveguide surface, without significant polymer desorption upon rinsing with buffer solution. The adsorbed mass of PLL-g-dex copolymers reported in Table 1 was measured, as well as that of dextran itself and PLL(20)g[3]-PEG(5), for comparison purposes: the results are presented in Figure 3. From the adsorbed polymer mass and the compositional features of the copolymers, it is possible to calculate the surface density of dextran chains, ndex, and lysine monomers, nlys, expressed as molecules nm-2, and the spacing between side chains on the surface, L, and finally to estimate the conformation of the surface-grafted dextran chains by comparing the spacing and the radius of gyration of dextran chains, L/2Rg. The results of these calculations are summarized in Table 2. All the PLL-g-dex copolymers investigated showed significant adsorption on the OWLS waveguides, ranging from ca. 200 to ca. 300 ng cm-2 on average, whereas dextran alone revealed negligible adsorption onto the surfaces (5.6 ( 3.7 ng cm-2). As with PLL-g-PEG copolymers, the adsorption of PLL-g-dex in an aqueous environment is thought to proceed through the electrostatic interactions between the polycationic PLL backbone and the surface: the dextran side chains stretch out toward the solution and the backbone lies flat on the substrate. In the case of dextran alone, no charges are available for electrostatic interactions, resulting in almost zero adsorption. The molecular weight (and therefore the length) of the PLL backbone and the density of the grafted dextran side chains along the backbone were the two architectural parameters changed for the synthesis of the PLL-g-dex copolymers employed in this work: 6 kDa and 13 kDa PLL backbones were used and the grafting ratio was systematically varied from g[3.9] to g[10.2]. Despite the variation in these two architectural parameters, the adsorbed masses obtained from the PLL-g-dex copolymers employed are nearly constant within the error bars, except for PLL(6)-g[10.2]-dex(5.9), which presumably possesses too few

8854 Langmuir, Vol. 24, No. 16, 2008

Perrino et al.

Table 2. Summary of the Adsorption Data Determined by OWLS for the PLL-g-dex Polymers.

a

surface

mpol (ng/cm2)

mserum

nLys (1/nm2)

ndex/PEG (1/nm2)

L (nm)

L/2Rg

PLL(13)-g[3.9]-dex(5.9) PLL(6)-g[5]-dex(5.9) PLL(13)-g[7.1]-dex(5.9) PLL(6)-g[10.2]-dex(5.9) PLL(20)-g(3)-PEG(5)

281 ( 12 281 ( 11 307 ( 58 217 ( 2 160 ( 3

13 ( 8 37 ( 4 17 ( 12 102 ( 20 13 ( 8

1.03 ( 0.05 1.29 ( 0.05 1.90 ( 0.39 1.85 ( 0.02 0.54 ( 0.00

0.26 ( 0.01 0.26 ( 0.01 0.27 ( 0.06 0.18 ( 0.00 0.18 ( 0.00

2.09 ( 0.05 2.11 ( 0.04 2.10 ( 0.22 2.52 ( 0.01 2.53 ( 0.00

0.39 ( 0.01 0.39 ( 0.01 0.39 ( 0.04 0.47 ( 0.00 0.45 ( 0.00

a mpol ) adsorbed polymer mass, mserum ) mass of serum adsorbed, nlys ) surface concentration of lysine monomers, ndex/PEG ) surface concentration of dextran or PEG, L ) spacing between grafted dextran or PEG chains, L/2Rg ) degree of overlap of dextran or PEG chains.

Figure 4. Lysine monomer density (nlys) and dextran chain density (ndex) as a function of grafting ratio (g) for the PLL(13)-g-dex(5.9) and PLL(6)g-dex(5.9) copolymers.

dextran side chains to establish a brush, and they showed no clear trend as the molecular weight of the PLL backbone or grafting ratio was changed. The negligible influence of the PLL molecular weight on the adsorption properties is thought to be due to the fact that the length of PLL backbones selected in this work, 6 or 13 kDa, is sufficiently short that a “flat-lying” conformation of PLL is readily achieved in both cases. Second, the negligible influence of the grafting ratio can be attributed to the opposing effects of the molecular weight of a single PLLg-dex copolymer and the probability of adsorption onto surfaces as a function of the grafting ratio. For instance, the grafting ratios g[3.9] and g[5] for the PLL(13)-g[3.9]-dex(5.9) and PLL(6)g[5]-dex(5.9) represent very different total numbers of grafted dextran chains on a single copolymer molecule, 16 and 5.7, respectively, and different total molecular weights, 88 and 28 kDa, respectively (structural details are summarized in Table 1). In terms of molecular weight, the copolymer with the lowest grafting ratio, g[3.9], might show higher mass of surface adsorption due to the higher number of dextran chains per single molecule. The number of anchoring groups (free lysine groups) is, however, higher for higher grafting ratios, while at the same time the steric hindrance between neighboring dextran side chains is smaller, both of which are advantageous for polymer adsorption. The control of the amount of polymer adsorption by the balance between the attractive electrostatic backbone-surface interactions and a steric repulsion between dextran side chains is more directly manifested in the plot of the lysine monomer density (nlys) and the dextran chain density (ndex) of the copolymers as a function of the grafting ratio (Figure 4). If we still consider the PLL(6)g[10.2]-dex(5.9) as an outlying exception, the lysine (or the polymer) molar surface density increases almost linearly with increasing grafting ratio, whereas ndex, which represents the efficiency of the copolymers in grafting dextran chains onto the surface, is nearly constant. Among the four copolymers investigated, PLL(6)-g[10.2]-dex(5.9) is the one with the highest

Figure 5. Adsorbed mass of serum, mserum, as a function of the surface concentration of grafted dextran or PEG chains, ndex/PEG.

grafting ratio, i.e., the lowest molecular weight; in this case the effect of the molecular weight overwhelms that of the available anchoring points and/or steric hindrance in determining the amount of polymer adsorbed at the equilibrium that a smaller adsorbed mass was observed. Compared to PLL(20)-g[3]-PEG(5), all PLL-g-dex copolymers examined showed higher adsorbed masses (200-300 ng cm-2 vs ca. 160 ng cm-2), although the molecular weight is definitely higher for PLL(20)-g[3]-PEG(5) than for all the PLL-g-dex copolymers. The difference in the radius of gyration (Rg) of the side chains, 2.68 nm for dextran(5.9)54 vs 2.82 nm for PEG(5),55 might explain the different amounts of polymer adsorbed: a smaller Rg of the side chains of the copolymer is, in fact, expected to enhance the surface adsorption due to the weaker shielding of the free lysine monomers (anchoring groups) by the side chains. In addition, the smaller the Rg, the weaker the steric interactions between neighboring side chains. QuantitatiVe Analysis of Protein Adsorption. OWLS measurements revealed that the amount of protein adsorbed can be greatly reduced by the presence of a PLL-g-dex adlayer on the surface. As summarized in Table 2 and shown in Figure 3, silica-titania surfaces exposed to a dextran solution, which showed negligible adsorption of dextran (5.6 ( 3.7 ng cm-2) and can be therefore considered as bare substrates, adsorbed a significant amount of serum (756 ( 34 ng cm-2), whereas all PLL-g-dex coated waveguides revealed a significant reduction in protein adsorption. Similarly to PLL-g-PEG, as shown in a previous study employing a broad range of polymer architectures (from ca. g[2] to g[20], PEG molecular weight ) 1, 2, and 5 kDa),21 the ability to resist nonspecific protein adsorption of the PLL-g-dex copolymers examined in this work was observed to be dependent on the surface density of the dextran chains (Figure 5): the amount of serum adsorbed decreases as the dextran density at the surface, ndex, increases. As long as the grafting density is sufficiently low (54) Go¨risch, S. M.; Wachsmuth, M.; Fejes Toth, K.; Lichter, P.; Rippe, K. J. Cell Sci. 2005, 118, 5825–5834. (55) Kawaguchi, S.; Imai, G.; Suzuki, J.; Miyahara, A.; Kitano, T.; Ito, K. Polymer 1997, 38(12), 2885–2891.

Nonspecific Adsorption of Proteins

Langmuir, Vol. 24, No. 16, 2008 8855

Figure 7. AFM results: friction force measured as a function of increasing load for the contact of a bare silicon nitride tip and silicon wafers coated with PLL-g-dex copolymers. Figure 6. Adsorbed mass of serum, mserum, as a function of the degree of overlap of dextran or PEG chains, L/2Rg. L denotes the average mean distance between dextran or PEG chains of the PLL-g-dex/PLL-g-PEG coated surfaces, calculated from the mass of polymer adsorbed and from the grafting ratios.

and thus a sufficiently high value for ndex is obtained, PLL-g-dex copolymers greatly reduce the nonspecific adsorption of proteins, reaching a level comparable to that observed for PLL-g-PEG coatings. The protein-resistance capabilities of PLL-g-PEG copolymers have previously proven to be dependent also on another factor: the degree of overlap between the PEG chains. This can be viewed as the product of the PEG chain length and density, and it is still related to the surface density of the surface-grafted chains.21 It can be quantitatively estimated by using a characteristic parameter, L/2Rg, which compares the average spacing between neighboring polymer chains, L, to their radius of gyration, Rg. In the case of L/2Rg being smaller than 1, the grafted polymer chains would overlap, were they to be in a mushroom configuration, and therefore they stretch out to form a brush structure. As shown in Figure 6, the dependence of the antifouling properties on L/2Rg was also observed for the PLL-g-dex copolymers investigated. According to what was observed in this work, however, dextran chains need to overlap more (L/2Rg e 0.4) than PEG chains (L/2Rg ) 0.4-0.7,21) to render the surface protein resistant. Compared to the PLL-g-PEG copolymer involved in this work, much higher values of surface density in the grafted chains and lower values of L/2Rg are needed for PLL-g-dex copolymers to achieve comparable antifouling capabilities, as judged from the adsorbed amount of proteins from serum (Figure 3 and Table 2). For example, PLL(6)-g[10.2]dex(5.9), which shows almost the same surface chain density and a similar value of L/2Rg to those of PLL(20)-g[3]-PEG(5), shows markedly inferior protein-repelling properties. This might be attributed to the higher flexibility and degree of hydration of PEG chains compared to dextran’s bulkier sugar units and, consequently, to the larger excluded volume of PEG in water, which allows PEG chains to form a more significant barrier against protein adsorption, even at smaller degree of side-chain overlap. Nanotribological Analysis. Nanotribological measurements of the PLL-g-dex copolymers by means of AFM served as a further probe of the difference in the architectural features of the copolymers. Previous studies have shown that the tribological

properties of graft copolymers, such as PLL-g-PEG, are significantly influenced by the architectural features, including side chain length,48,56 grafting ratio,48 and the backbone length.56 The nanotribological properties characterized by AFM (Figure 7) are indeed able to clearly distinguish between copolymers with different architectural features (grafting ratio and length of backbone), most of which were indistinguishable in their adsorption behavior and the protein-resistance properties. For instance, the two copolymers with 13 kDa PLL backbone and PLL(6)-g[5]-dex(5.9), which were only marginally different in their surface adsorption and antifouling properties, were observed to reveal distinctively different frictional behavior; the frictional forces being in the order of PLL(6)-g[10.2]-dex(5.9) . PLL(13)g[3.9]-dex(5.9) > PLL(6)-g[5]-dex(5.9) > PLL(13)-g[7.1]dex(5.9). Significantly higher frictional forces observed from the PLL(6)g[10.2]-dex(5.9) can be attributed to the distinctly smaller surfacegrafted dextran chain density, ndex (Table 2), which, in turn, leads to its inferior ability to generate an aqueous lubricating film at the interface. On the other hand, since the ndex and L/2Rg values for the other copolymers are fairly similar, the improved lubricating behavior with increasing grafting ratio, from 3.9 to 7.1, might be associated with the stability of the polymer adlayer. For instance, the copolymer with higher grafting ratio, such as PLL(13)-g[7.1]-dex(5.9), may have stronger binding to the surface due to the higher number of available anchoring points, as well as lower steric hindrance between side chains, allowing easier access to the surface by the positively charged amine groups on the backbone. Finally, as with the surface adsorption and antifouling properties, the effect of the molecular weight of the PLL backbone between 6 and 13 kDa was observed to be inconsequential for the nanotribological properties.

Conclusions The protein-repelling capabilities of a series of dextran-based graft copolymers (PLL-g-dex) with different grafting ratios and lengths of the PLL backbone, were investigated by means of OWLS and compared to PLL-g-PEG, a graft copolymer with the same PLL backbone but with PEG as side chains, which has been shown to be highly effective at preventing nonspecific protein adsorption. Similarly to PLL-g-PEG, the amount of proteins adsorbed onto PLL-g-dex-coated silica-titania waveguides was observed to be dependent on the surface concentration (ndex) and degree (56) Mu¨ller, M.; Lee, S.; Spikes, H. A.; Spencer, N. D. Tribol. Lett. 2003, 15(4), 395–405.

8856 Langmuir, Vol. 24, No. 16, 2008

of overlap (L/2Rg) of the surface-grafted polymer chains: PLLg-dex copolymers with high values of ndex (g0.26 nm-2) greatly reduce the nonspecific adsorption of proteins to a level comparable to those observed from PLL-g-PEG coatings. Nanotribological measurements by AFM were performed in order to further distinguish between the investigated copolymers: Those PLLg-dex copolymers that showed the optimum, and yet indistinguishable, surface adsorption and antifouling properties resulted in notably different lubricating properties. For the PLL-g-dex copolymers employed in this work, the grafting ratio was observed to be the most important architectural parameter in determining

Perrino et al.

the nanotribological properties. While the tribological studies in this work were motivated primarily by an attempt to better probe the architectural features of the PLL-g-dex copolymers, the combination of good lubricating and antifouling properties is highly desirable for some biomedical applications involving moving parts, e.g., coatings for stents, contact lenses, catheters, and endoscopes. The comparison of OWLS and AFM data can in this case be very helpful for the selection of the “ideal” polymer architecture for such applications. LA800947Z