Langmuir 2006, 22, 313-318
313
Colloid Probe AFM Investigation of Interactions between Fibrinogen and PEG-Like Plasma Polymer Surfaces Kristen E. Bremmell,*,† Peter Kingshott,‡ Zahida Ademovic,‡ Bjørn Winther-Jensen,‡ and Hans J. Griesser† Ian Wark Research Institute, UniVersity of South Australia, Mawson Lakes, SA 5095, Australia, and Danish Polymer Centre, Risø National Laboratory, FrederiksborgVej 399, 4000 Roskilde, Denmark ReceiVed August 5, 2005. In Final Form: October 23, 2005 Interaction forces between surfaces designed to be protein resistant and fibrinogen (Fg) were investigated in phosphatebuffered saline with colloid probe atomic force microscopy. The surfaces of the silica probes were coated with a layer of fibrinogen molecules by adsorption from the buffer. The technique of low-power, pulsed AC plasma polymerization was used to make poly(ethylene glycol) (PEG)-like coatings on poly(ethylene teraphthalate) by using diethylene glycol vinyl ether as the monomer gas. The degree of PEG-like nature of the films was controlled by use of a different effective plasma power in the chamber for each coating, ranging from 0.6 to 3.6 W. This produced a series of thin films with a different number of ether carbons, as assessed by X-ray photoelectron spectroscopy. The interaction force measurements are discussed in relation to trends observed in the reduction of fibrinogen adsorption, as determined quantitatively by 125I radio-labeling. The plasma polymer coatings with the greatest protein-repelling properties were the most PEG-like in nature and showed the strongest repulsion in interaction force measurements with the fibrinogen-coated probe. Once forced into contact, all the surfaces showed increased adhesion with the protein layer on the probe, and the strength and extension length of adhesion was dependent on both the applied load and the plasma polymer surface chemistry. When the medium was changed from buffer to water, the adhesion after contact was eliminated and only appeared at much higher loads. This indicates that the structure of the fibrinogen molecules on the probe is changed from an extended conformation in buffer to a flat conformation in water, with the former state allowing for stronger interaction with the polymer chains on the surface. These experiments underline the utility of aqueous surface force measurements toward understanding protein-surface interactions, and developing nonfouling surfaces that confer a steric barrier against protein adsorption.
Introduction Interaction of plasma proteins with the material surface dictates the response of the body to cardiovascular implants. The nonspecific adsorption of proteins is known to lead to complications. Therefore, the development of protein- and cell-repellent biocompatible surfaces has received significant effort in biomaterials engineering research. Poly(ethylene oxide) (PEO), also referred to as poly(ethylene glycol) (PEG) has been shown to be effective in reducing protein and cell adhesion in a number of studies1-9 and is compatible with living cells.10 Mechanisms of protein resistance have been a frequently discussed topic in the literature. Poly(ethylene glycol) (PEG)modified surfaces are thought to have excellent repellent properties attributable to steric stabilization by the flexible PEG chains.1-5,7,11,12 It has been reported that PEG chains do not * Corresponding author. E-mail:
[email protected]. † Ian Wark Research Institute, University of South Australia. ‡ Danish Polymer Centre, Risø National Laboratory. (1) Amiji, M.; Park, K. J. Biomater. Sci. Polym. 1993, Ed. 4, 217. (2) Elbert, L.; Hubbell, J. A. Annu. ReV. Mater. Sci. 1996, 26, 365. (3) Halperin, A. Langmuir 1999, 15, 2525. (4) Harris, J. M. Poly(ethylene glycol) Chemistry: Biotechnical and Biomedical Applications; Plenum Press: New York, 1992. (5) Israelachvili, J. Proc. Natl. Acad. Sci. U.S.A. 1997, 94, 8378. (6) Lee, J. H.; Lee, H. B.; Andrade, J. D. Prog. Polym. Sci. 1995, 20, 1043. (7) Kingshott, P.; Thissen, H.; Griesser, H. J. Biomaterials 2002, 23, 2043. (8) Kingshott, P.; McArthur, S.; Thissen, H.; Castner, D. G.; Griesser, H. J. Biomaterials 2002, 23, 4775. (9) Pasche, S.; Textor, M.; Meagher, L.; Spencer, N. D.; Griesser, H. J. Langmuir 2005, 21, 6508. (10) Albertsson, P.-A° . Partition of Cell Particles and Macromolecules, 3rd ed.; Wiley: New York, 1986. (11) Jeon, S. I.; Lee, J. H.; Andrade, J. D.; de Gennes, P. G. J. Colloid Interface Sci. 1991, 142, 149. (12) Jeon, S. I.; Andrade, J. D. J. Colloid Interface Sci., 1991, 142, 159.
interfere with the hydrogen-bonded water structure significantly,13 a common factor seen in surfaces that show nonspecific adsorption of biomolecules. Additionally, in aqueous solution, PEG molecules are highly mobile.14 As a result of the significant interest in PEG-modified surfaces, a number of approaches to preparing PEG-coated surfaces have been developed and examples in the literature are numerous; for example, studies involving physical adsorption of copolymers,9,15-17 chemical grafting,7,16,18-20 and through adsorption of supported lipid bilayers displaying end-grafted PEG.21,22 Despite various studies of protein and cell adsorption, there remains controversy in the results and as to whether the length of the PEG chain or the density on the surface dominate the macromolecular adsorption/adhesion, though recent experiments suggest that suitable combinations of the two factors are required.7,9 An alternative approach has been to deposit PEG-like coatings by plasma deposition using volatile monomers containing ethylene (13) Kitano, H.; Kawasaki, A.; Kawasaki, H.; Morokoshi, K. J. Colloid Interface Sci. 2005, 282, 340. (14) Nagaoka, S.; Mori, Y.; Takyuchi, H.; Yokota, K.; Tanzawa, H.; Nishiumi, S. In Polymers as Biomaterials; Shalaby, S. W., Hoffman, A. S., Ratner, B. D., Horbett, T. A., Eds; Plenum Press: New York, 1985; p 361. (15) Musoke, M.; Luckham, P. E. J. Colloid Interface Sci. 2004, 277, 62. (16) Norde, W.; Cage, D. Langmuir 2004, 20, 4162. (17) Pasche, S.; De Paul, S. M.; Vo¨ro¨s, J.; Spencer, N. D.; Textor, M. Langmuir 2003, 19, 9216. (18) Archambault, J. G.; Brash, J. L. Colloids Surf., B. 2004, 39, 9. (19) Drumheller, P. D.; Herbert, C. B.; Hubbell, J. A. Bioprocess Technol. 1996, 23, 273. (20) Mougin, K.; Lawrence, M. B.; Fernandez, E. J.; Hillier, A. C. Langmuir 2004, 20, 302. (21) Sheth, S. R.; Leckband, D. Proc. Natl. Acad. Sci. U.S.A. 1997, 94, 8399. (22) Sheth, S. R.; Efremova, N.; Leckband, D. J. Phys. Chem. B. 2000, 104, 7652.
10.1021/la052143a CCC: $33.50 © 2006 American Chemical Society Published on Web 11/17/2005
314 Langmuir, Vol. 22, No. 1, 2006
oxide units (-CH2CH2O-).23-29 As plasma deposition is a complex process involving monomer fragmentation, rearrangement, cross-linking, and polymerization,30 the resulting polymer coating chemistry is different from that of polymers formed by conventional polymerization methods. It has been shown in previous studies that the molecular structure of plasma-deposited PEG-like films is randomly cross-linked and consists of methylterminated ethylene oxide chains.23,26 In addition, through control of the power supplied during plasma polymerization, the ether carbon content of the PEG-like coatings can be changed,27 and when the monomer structure was retained in the polymerized film, it was shown to remain hydrogel-like in nature. The molecular mechanisms of interaction of proteins with surfaces has inspired a number of studies.16,21,31-34 Atomic force microscopy (AFM) has been used to explore interactions between polymer-coated surfaces15 and protein interactions with surfaces.15-17,21,22,26,31-33,35 McGurk et al.32 investigated the interaction of an albumin-coated sphere with various poly(ethylene oxide)-polypropylene oxide triblock polymer (PEOPPO-PEO)-coated surfaces and found the adhesion forces to correlate with adsorption of the protein measured by surface plasmon resonance (SPR). Gergely et al.34 found that the mean rupture force between fibrinogen (Fg) adsorbed on the cantilever tip and a silica surface increased with the loading rate. Further, several ruptures were observed for each approach-and-retract cycle, interpreted as corresponding to the detachment of successive domains of protein molecules from the surface. Alternative techniques have also demonstrated that some proteins can interact with PEO.36-40 For example, it has been reported that PEG binds to hydrophobic patches of proteins under certain conditions to stabilize partially unfolded proteins against aggregation38 and that PEG forms noncovalent complexes with chymotrypsin under pressure and at elevated temperatures (where interactions of the ethylene oxide group with water are weakened).40 Investigation of the interaction between the plasma protein fibrinogen and PEG-like surfaces prepared with different applied current during plasma polymer deposition of diethylene glycol vinyl ether (DEGVE) is the focus of the present study. XPS surface characterization, radio-labeled protein adsorption studies, and direct AFM interaction force measurements between (23) Lopez, G. P.; Ratner, B. D.; Tidwell, C. D.; Haycox, C. L.; Rapoza, R. J.; Horbett, T. A. J. Biomed. Mater. Res. 1992, 26, 415. (24) Lo´pez, G. P.; Ratner, B. D. Plasmas Polym. 1996, 1, 127. (25) Johnston, E. E.; Bryers, J. D.; Ratner, B. D. Langmuir 2005, 21, 870. (26) Pan, Y. V.; McDevitt, T. C.; Kim, T. K.; Leach-Scampavia, D.; Stayton, P. S.; Denton, D. D.; Ratner, B. D. Plasmas Polym. 2002, 7, 171. (27) Beyer, D.; Knoll, W.; Ringsdorf, H.; Wang, J.; Timmons, R. B.; Sluka, P. J. Biomed. Mater. Res. 1996, 36, 181. (28) Ademovic, Z.; Wei, J.; Winther-Jensen, B.; Hou, X.; Kingshott, P. Plasma Polym. Proc. 2005, 2, 53. (29) Wu, Y. L. J.; Timmons, R. B.; Ten, J. S.; Molock, F. E. Colloids Surf., B. 2000, 18, 235. (30) Yasuda, H. Plasma Polymerisation; Academic Press: Orlando, FL, 1985. (31) Meagher, L.; Griesser, H. J. Colloids Surf., B. 2002, 23, 125. (32) McGurk, S. L.; Green, R. J.; Giles, H. W.; Davies, M. C.; Roberts, C. J.; Tendler, S. J. B. Langmuir 1999, 15, 5136. (33) Zhang, H.; Bremmell, K. E.; Smart, R. St. C. J. Biomed. Mater. Res. A. 2005, 74, 59. (34) Gergely, C.; Voegel, J.-C.; Schaaf, P.; Senger, B.; Maaloum, M.; Ho¨rber, J. KH.; Hemmerle´, J. Proc. Natl. Acad. Sci. U.S.A. 2000, 7, 10802. (35) Rixman, M. A.; Dean, D.; Oritz, C. Langmuir 2003, 19, 9357. (36) Abbott N. L.; Blankschtein, D.; Hatton, T. A. Macromolecules 1992, 25, 3932. (37) Azegami, S.; Tsuboi, A.; Izumi, T.; Hirata, M.; Dubin, P. L.; Wang, B.; Kokufuta, E. Langmuir 1999, 15, 940. (38) Cleland, J. L.; Wang, D. I. C. Biotechnology 1993, 3, 527. (39) Efremova, N. V.; Bondurant, B.; O’Brien, D. F.; Leckband, D. E. Biochemistry 2000, 39, 3441. (40) Topchieva, I. N.; Sorokina, E. M.; Efremova, N. M.; Ksenofontov, A. L. Biochemistry (Moscow) 1998, 63, 1312.
Bremmell et al.
fibrinogen and the surfaces have been utilized to characterize the protein-resistant nature of the modified surfaces. Materials and Methods Coatings and Chemicals. Poly(ethylene teraphthalate) (PET) film of 150 µm thickness (Traforma A/S, Denmark) was cut into 13 mm diameter disks and cleaned by Soxhlet extraction with hexane/ ethanol. Phoshpate-buffered saline (PBS) and DEGVE were purchased from Sigma-Aldrich (Germany). Unlabeled fibrinogen (Fg) was obtained from Calbiochem, and 125I radio-labeled fibrinogen (125I-Fg) was purchased from Amersham Bioscience. Plasma Deposition. Plasma coating was carried out in a custombuilt plasma unit as described previously.28,41 The plasma chamber is a conventional cylindrical stainless steel vacuum chamber with an actual plasma volume of 11.5 L. The stainless steel electrodes (0.5 mm thickness) are isolated from the vacuum chamber. Prior to deposition, the samples were pretreated with an Ar plasma for surface activation and to improve adhesion between the substrate and the plasma polymer coating. We believe the coatings are between 50 and 100 nm thick and form a continuous film over the PET surface. Surface Characterization. X-ray photoelectron spectra (XPS) were recorded by using a SSX-100 spectrometer (Surface Science Laboratories, USA) equipped with an Al KR source at a power of 150 W. Total pressure in the main chamber was typically 1 × 10-8 bar. The emission angle was set at 55° with respect to the sample normal, which results in an information depth of approximately 6 nm. A low-energy electron flood gun at an energy of 6 eV was used for charge compensation. Curve fitting was performed (Spectra Data processor version 2.3 software) by using linear background subtraction. The binding energy of the CHx C 1s component at 285.0 eV was used as the reference for charge correction. Protein Adsorption. Fibrinogen (150 µg/mL) was prepared in a citrated PBS (cPBS) buffer at pH 7.4 that contained 10 mM citric acid, 10 mM Na2HPO4, and 120 mM NaCl. 125I-labeled fibrinogen was added to adjust to a specific activity of at least 2000 cpm/µg protein. The unlabeled/labeled ratio was 20:1. Unmodified and coated PET disks were incubated in a solution of fibrinogen for 1 h at 37 °C, prior to rinsing 4 times with cPBS. A Canberra 20 gamma counter (Canberra, USA) was used to measure the radioactivity of the disks, and the amount of adsorbed fibrinogen was calculated. The background counts were 35 cpm in the 125I window. Atomic Force Microscopy. An atomic force microscope (Nanoscope III, Veeco) was used to measure direct interaction forces between a fibrinogen-coated glass sphere and the modified polymer surfaces in PBS and Milli-Q water. A silica glass sphere (Alltech), 5 µm in diameter, was glued to a silicon nitride V-shaped cantilever (DI, USA) by using Epikote heat-sensitive resin and then soaked in a solution of 1 M NaOH, rinsed with Milli-Q water, and plasma cleaned for 20 s. When required, the cantilever with attached silica sphere was conditioned in a solution of fibrinogen for 1 h for protein adsorption to take place. ToF-SIMS imaging was performed on the fibrinogen-coated silica sphere by using a Physical Electrics TRIFT instrument (data not published) before and after one set of experiments. This indicted the presence of amino acid fragments covering the surface, indicating that the protein remains adsorbed to the probe during the experiments. In addition, the same probe was used for all measurements shown in this study, and when repeating measurements on the PET control surface and on the first plasma polymer surface used, similar force curves were obtained, indicating that the protein did not desorb or change interaction behavior with the surface after repeated measurements. The cantilever spring constants were measured by using the technique of Cleveland et al.,42 and values in the range of 0.09-0.10 N/m were determined. By being operated in a fluid cell, the interactions were measured in PBS and Milli-Q water after a conditioning period of 15 min for each solution. The data were exported to Excel and transformed into (41) Glejbøl, K.; Winther-Jensen, B. (NKT Research A/S). U.S. Patent 6,628,084, 2003. (42) Cleveland, J. P.; Manne, S.; Bocek, D.; Hansma, P. K. ReV. Sci. Instrum. 1993, 64, 403.
Fibrinogen and PEG-Like Plasma Polymer Surfaces
Langmuir, Vol. 22, No. 1, 2006 315
Table 1. XPS Results for Elemental Composition, Binding Energies, and Components of the Carbon C1s Signal for Untreated PET and for DEGVE-Coated PET Surfaces C1s atom. % surface
C-C, C-H 285 eV
C-O, C-N 286.5 eV
CdO 288.0 eV
O-CdO 289.4 eV
O1s atom. %
PET PET-DEGVE (2.5 mA) PET-DEGVE (5 mA) PET-DEGVE (10 mA)
38.5 25.5 33.4 34.9
16.3 40.4 34.5 29.7
5.3 6.1 7.1
17.1 0.4 0.3 2.0
28.0 28.4 25.7 26.3
Table 2.
125
I-Fibrinogen Adsorption to PET and the Plasma-Coated PET Surfaces
sample
Fg (adsorbed) [ng cm-2]
PET PET + DEGVE (2.5 mA) PET + DEGVE (5 mA) PET + DEGVE (10 mA)
881 ( 112 92 ( 13 178 ( 14 218 ( 39
force/radius vs apparent separation plots. All separations are relative to the distance between the surfaces when the system had attained constant compliance. The polymer layer is thus pressed between the silica sphere and the PET substrate, and the distance of closest approach is thus not known exactly. Here, we present forces scaled by the effective radius of the sphere, R, to relate the force, F, between a curved and a flat surface to the interaction free energy, E, between parallel flat plates of unit area by eq 1.43 F/R ) 2πE
(1)
Results PET substrates were modified with DEGVE by pulsed AC plasma polymerization at different plasma powers to produce coatings that vary in their similarity to PEG. The surface chemistry of such films has been previously characterized extensively with XPS, IR, and ToF-SIMS.28 XPS results of coatings used in this study are shown in Table 1 and demonstrate significant variations in the chemical nature of the modified surfaces with respect to the power applied during the plasma polymer deposition. The PEG-like nature of the surface can be associated with the relative contribution of the ether/alcohol peak (C-O/C-OH) at a binding energy of 286.5 eV to the total C 1s signal of the XPS spectra. The C-O/C-C ratio decreases as the power applied in deposition increases. This implies a greater retention of ethylene oxide monomer structure in the films produced with lower powers. Changes in the surface chemistry were also confirmed with ATRIR spectroscopy.28 Characteristic peaks of the aliphatic ether at approximately 1150 cm-1 and the carbonyl peak (∼1720 cm-1) were observed. As the plasma power used in the DEGVE deposition was decreased, the ratio of the ether to carbonyl content was observed to increase, further indicating that the ether structure is increasingly retained by use of a lower plasma power. Further detailed discussion is included in an earlier publication.28 The capacity of these modified surfaces to resist protein adsorption was investigated by using 125I radio-labeled fibrinogen.28 In Table 2, the amount of Fg adsorbed to the coated surfaces is shown. Fg adsorbs to approximately a monolayer on the unmodified PET surface. After surface modification with DEGVE, protein adsorption was reduced significantly and was observed to increase with applied plasma power. On the DEGVE surface deposited using the lowest plasma current of 2.5 mA, less than 95 ng cm-2 Fg adsorbed, which corresponds to a 90% reduction in protein adsorption. As the plasma current was increased to 10 mA, the protein adsorption increased to more (43) Israelachvili, J. N. Intermolecular and Surface Forces, 2nd ed.; Academic Press: London, 1991.
Figure 1. Interaction forces, normalized by the probe radius, on approach between a silica sphere and the DEGVE plasma-coated surfaces in PBS.
than 200 ng cm-2. Therefore, the ability of the DEGVE surfaces to reduce protein adsorption is directly related to the PEG-like nature of the surface modification observed in the XPS data. To characterize Fg interactions with the DEGVE-modified surfaces, and attempt to correlate the results of the Fg quantification, analysis of the surface forces of DEGVE coatings and interaction forces with Fg was performed by colloid probe microscopy. Prior to adsorbing Fg to the colloid probe surface, the interaction between a bare silica sphere and the DEGVEmodified surfaces in PBS was examined. Figure 1 illustrates the direct interaction forces measured when the silica sphere approached each of the surfaces in PBS. This behavior was reproducible in more than 10 successive measurements and was observed at several different regions of the sample surface. For all DEGVE-modified surfaces, a repulsive force was observed. The decay length of the interaction force decreased with increasing plasma power (or current) used to generate the DEGVE coatings (the decay lengths were calculated to be 1.70 ( 0.34 nm for DEGVE 10, 3.04 ( 0.30 nm for DEGVE 5, and 4.81 ( 0.35 nm for DEGVE 2.5). However, these are too long for the interaction to result from electrical double-layer overlap (in PBS, κ-1 ) 0.8 nm) and suggests that the repulsion is derived from a steric interaction. This implies that the surface prepared at the lowest current (DEGVE 2.5) has the greatest steric repulsive interaction with the silica sphere. Once the surfaces had been pressed into contact, the interaction on separation was measured (Figure 2). In PBS, the silica sphere adhered to all surfaces. Adhesion between the silica colloid probe and DEGVE-modified surfaces is not surprising, as PEO is known to adsorb on inorganic oxide surfaces, particularly silica.44,45 The adhesion observed decreases both in strength and extension distance with increasing plasma deposition power. As the level of cross-linking and fragmentation increases with higher power used during deposition, adhesion with the silica sphere decreases, probably as a consequence of the presence of shorter, less flexible PEO-like segments in higher power coatings. Polymer segments probably need to adapt and adhere to the colloid surface with a number of units in order to produce observable adhesion. (44) Braithwaite, G. J. C.; Howe, A.; Luckham, P. F. Langmuir 1996, 12, 4224. (45) Mathur, S.; Moudgil, B. M. J. Colloid Interface Sci. 1997, 196, 92.
316 Langmuir, Vol. 22, No. 1, 2006
Bremmell et al.
Figure 2. Normalized interaction forces measured on separation after contact between the silica sphere and DEGVE-modified surfaces in PBS. Figure 4. Normalized interaction forces measured on separation of the Fg-coated silica sphere and PET and DEGVE plasma-coated surfaces after contact. The extent of the adhesive events decreases for DEGVE 10.
Figure 3. Interaction forces, normalized by the probe radius, measured on approach between a fibrinogen-coated sphere and PET and DEGVE plasma-coated surfaces in PBS.
To measure Fg interactions with the DEGVE surfaces, silica spheres attached to cantilevers were exposed to a Fg solution (in PBS), prior to insertion in the AFM fluid cell, under conditions that would ensure that the surface would acquire monolayer coverage of the protein (assumed from our quantitative measurements of adsorption by using 125I-labeled Fg on glass microscope slides and silicon wafers, where ∼650 ng/cm2 of Fg adsorbs, which is equivalent to a mixed monolayer of molecules adsorbed in random configuration). Direct interaction forces between the protein-covered probe and all the DEGVE surfaces as well as the PET control, are shown in Figure 3. Initially, for the unmodified PET surface, the Fg-coated silica sphere experiences a strong attraction with the surface in PBS. In contrast, when the fibrinogen-coated sphere approaches the DEGVE-modified surfaces, a repulsive interaction force is observed. In PBS, the electrical double-layer interactions are screened (κ-1 ) 0.8 nm), and after fitting the curves with linear Poisson-Boltzmann theory, the repulsion is 4 times larger than expected and the repulsive force thus is a result of steric repulsion between the Fg and the DEGVE film. In addition, the repulsive force curve on approach did not alter with ionic strength. Thus, the repulsion observed does not arise from electrical double-layer overlap and implies that the force is due to steric interactions between the protein and DEGVE plasma polymer surfaces. The extent of the steric repulsion is different for the three DEGVE surfaces. For the DEGVE 2.5- and 5-modified surfaces (prepared at 2.5 and 5 mA, respectively), the steric repulsion is of similar magnitude and extent. However, for the polymer film generated with 10 mA plasma current, the magnitude and extent of the repulsive force is observed to decrease. This suggests that the steric
interactions responsible for repulsion of the Fg-coated sphere have been reduced. At the higher plasma power used to generate the DEGVE 10 polymer film, more monomer fragmentation will take place, which results in the less hydrocarbon nature of the film (as observed in XPS). In accordance with the chemical surface analysis and higher Fg adsorption results, the DEGVE38 plasma polymer surface may be interpreted as being more crosslinked and dense, with fewer protruding linear chains as the power used during DEGVE film preparation was increased. Once the surfaces have been pressed into contact and have thus overcome the repulsive force, an adhesion between the surfaces is often observed. Typical interaction forces measured on separation of the Fg-coated sphere and the polymer surface in PBS are illustrated in Figure 4. When the silica Fg-coated sphere interacts with PET, the adhesion experienced is too large to measure accurately within the experimental conditions, and on release of the probe from the surface, the separation between the probe and the surface is 120 nm. For the DEGVE-coated surface, several rupture events were observed in each retraction curve. In previous studies, these rupture events were assigned to the detachment of successive domains of proteins.34 In PBS, the separation distance that the adhesive events extend to decreases with plasma power. The magnitude of adhesion also decreases (calculated from 10 different force curves, the maximum adhesion observed for the 2.5 mA was 0.063 ( 0.01 mN/m, for 5 mA, 0.043 ( 0.003, and for the 10 mA surface, 0.033 ( 0.008 mN/ m). Adhesion between proteins and PEO-like surfaces has been observed in other studies21,35 and for chemically modified surfaces.22 Sheth and Leckband21 measured the interaction between streptavidin and a PEO-modified surface by using the surface force apparatus (SFA). A repulsion was observed at long range; however, when forced closer together, the interaction became attractive. The attraction was sufficiently strong to produce an adhesion between the protein and PEO-coated surfaces. Similarly, Rixman et al.35 measured attractive contacts between human serum albumin and PEO-modified surfaces by using high-resolution force spectroscopy. When increasing the load applied between the Fg-coated probe and the surfaces, the interaction force curves in Figure 5 a, b, and c show a change in adhesive interaction forces upon separation of the Fg-coated sphere with the DEGVE-modified surfaces. There is an overall trend of the adhesion increasing in magnitude and in the number of adhesive events with increasing load. As the load is increased, the time the surfaces spend in contact and
Fibrinogen and PEG-Like Plasma Polymer Surfaces
Langmuir, Vol. 22, No. 1, 2006 317
Figure 6. Interaction force, normalized by the probe radius, measured between a Fg-coated sphere and DEGVE 5 in water (red dots) and in PBS.
occurred even at low applied loads and did not vary with load as significantly as for the other two, more PEG-like surfaces. Changing solution conditions was also observed to impact on the adhesion behavior between Fg and the DEGVE-modified surfaces. Figure 6 shows that there is no interaction between the Fg and DEGVE 5 surface when PBS is replaced by water in the fluid cell. The interaction force is also reversible; replacing water by PBS again re-establishes the adhesive interaction (data not shown). Evidence from light scattering data has indicated that the dimension of the Fg molecule is reduced in water to 25 nm diameter, compared with 31 nm in PBS. It was also noted that Fg is less soluble in water than when ions are present in the solution (i.e., PBS or 0.1 PBS/water).
Discussion
Figure 5. Direct interaction forces, normalized by the probe radius, measured on separation of a Fg-coated silica sphere and (a) a DEGVE 2.5 coating as a function of load, increasing from 0.044 to 0.25 mN/m, (b) a DEGVE 5 coating as a function of load, increasing from 0.041 to 0.19 mN/m, and (c) DEGVE 10 as a function of applied load from 0.07 to 0.12 mN/m.
the area of interaction will increase. It would, therefore, be expected that the average number of contacts between protein molecules and the polymer surface will increase. It was observed that the influence on load is greater for the surfaces produced by using lower power (Figure 5a and b). For these surfaces, at low loads, the Fg-coated sphere did not adhere to the surface, but as the load increased, the adhesion increased. However, for the surface prepared with the highest power, DEGVE 10, adhesion
Interaction force measurements between the silica sphere and DEGVE-modified surfaces (Figures 1 and 2) demonstrate that surface modification using plasma deposition with different powers has altered the physical nature of the surface to generate a repulsive steric character that increases when lower deposition powers are used. In previous studies, Musoke and Luckham15 observed that the repulsion between surfaces with adsorbed PEO-PPO-PEO copolymers became longer ranged as the molecular weight of the PEO chain increased, and Sheth et al.21 found the range and magnitude of the repulsion to increase with grafting density. Therefore, the results obtained for the modified surfaces in this study suggest that the surfaces have longer, or a greater concentration of, PEO-like chains when prepared with a lower power, both of which accord with the greater ethylene oxide content observed in the XPS data (Table 1). Correspondingly, Fg interaction with the modified surfaces shows the largest repulsion for the coatings that possess the largest steric barrier to the approach of protein molecules. Once the surfaces have been pressed into contact with sufficient force, adhesion was often observed of Fg to the plasma coating. Therefore, in the current system, interaction force measurements suggest that the protein resistance of PEG-coated surfaces appears to result from an activation energy against entering the surface layer, as opposed to repulsion between the protein and PEG-like surface. This is in agreement with previous studies, where force measurements indicated a repulsion at long distances between PEG brushes and streptavidin but an attractive contribution at closer range,21 and a weak attraction between BSA and PEO coils was observed in an aqueous two-phase system.46 Abbott et al.36,46 found that both statistical thermodynamics and structural models predict that an attractive interaction of approximately 0.05 kT (per polymer segment at the protein surface) exists (46) Abbott, N. L.; Blankschtein, D.; Hatton, T. A. Macromolecules 1991, 24, 4334.
318 Langmuir, Vol. 22, No. 1, 2006
Figure 7. Histogram showing the inter-rupture distances on separation of the Fg-coated sphere and polymer surface. The interrupture distances decrease as the plasma power used in polymerization increased.
between BSA and PEO. Sheth et al.22 studied the interaction between PEO and selectively chemically modified surfaces to characterize the interaction with different functional groups of proteins. They concluded that protein association with PEG surfaces may be mediated by interactions between polymer segments and apolar amino acid side chains. After compression of the Fg to the plasma coatings, adhesion occurred on separation, resulting in a number of pull-off events. The distance between the elastic events was determined for a number of force curves measured with similar applied loads and is illustrated in the histogram in Figure 7. The polymer surface prepared with the lowest power has rupture events extending up to 90 nm of separation. As the power used to produce the polymer films increases, the rupture distances measured decrease in extent. The DEGVE 10 surface produces adhesive events out to only 30 nm. These behaviors correlate with the reduction in repulsion observed between the surfaces and suggests that, as the surface polymer film becomes more rigid (DEGVE 10) and less PEGlike in nature, less entanglement or interaction between the protein and surface occurs. The modified surface using the highest power (DEGVE 10) is characterized by a lower concentration of ether groups and, therefore, is likely to interact with the protein by a different mechanism. While plasma polymer coatings are often thought to be extensively cross-linked and able to undergo only relatively low extents of swelling in solvents, the DEGVE plasma coatings of the present study are likely to be quite different. We believe that they are structurally more akin to hydrogel polymers, although we have not yet been able to measure their hydration. Particularly at low power, they probably form linear polymeric, PEO-like structures, and the contour length of such segments is reflected in the observed pull-off events of Figure 2, against a silica sphere probe. When Fg is added to the probe, the distance dependence of the pull-off events (Figures 4 and 5) is a superposition of stretching of polymer chain segments and of Fg segments. It is interesting to consider the effect of changing the load applied between the Fg-coated sphere and the plasma coatings. This series of experiments suggests that a threshold force is required to overcome the repulsive steric interaction between the Fg and DEGVE surfaces prepared at low power inputs for the attractive adhesion to occur on separation. Thus, the Fg-
Bremmell et al.
coated sphere needs to be substantially pushed into the DEGVE plasma coating to initiate adhesive interaction; the approach force evidently overcoming the entropic repulsion of the steric barrier layer. The energy barrier for the DEGVE plasma coating produced with higher power is lower, and therefore, adhesion between the protein and the coating surface occurs under lower applied loads. The results obtained in this study suggest that surfaces offering higher steric repulsion to hinder approach and adsorption by protein molecules will have a stronger “nonfouling” character. It has been discussed previously5 that if a short-ranged attraction exists, even in the presence of a longer-ranged repulsion, the system will remain stable only for a certain period of time; the system is kinetically stable, but not thermodynamically stable. Finally, the extent of interaction between Fg and PEG-like surfaces upon separation can be varied by changing the solution from PBS to water. Here it is observed that adhesion between Fg and the DEGVE 5 surface is removed when the solvent is replaced with water. A recent study9 has shown that increasing the solution ionic strength does not alter the steric repulsive forces observed between a dense covering of PEG chains and a silica sphere. Thus, when the solvent is replaced with water, it appears that the protein conformation and behavior is influencing the interactions observed here. Milli-Q water has a pH of 5.5, which corresponds to the pI of Fg. Therefore, the protein is less soluble. Fg is also likely to undergo denaturation in Milli-Q water; the resultant extended confirmation will adsorb more tightly on the probe surface and reduce the availability of protruding segments that can become involved in bridging. Only when very high loads are applied is there a weak interaction.
Conclusion Interactions between fibrinogen and plasma-polymerized DEGVE-modified PET surfaces have been investigated. The more “PEG-like” the nature of the surface is, the greater is the reduction in protein adsorption. Direct interaction force measurements between a protein-coated sphere and the modified surfaces in PBS demonstrate that there is a correlation between the adsorbed amount of fibrinogen and the repulsive force measured between the Fg-coated probe and DEGVE-modified surfaces. Enhanced repulsive interaction was observed when the surfaces became more “PEG-like” in nature. Once the protein has been forced into contact with the surface, the forces on retraction demonstrate multiple adhesive events that are longer ranged for surfaces with a greater PEG character. The load applied to the protein-coated probe when interacting with the surfaces determines the extent of adhesion. However, surfaces with a higher repulsive barrier to approach show no adhesion at low applied loads. Acknowledgment. We acknowledge an Ian Wark Research Institute Visiting Fellowship to Peter Kingshott to facilitate this collaboration. This work was also supported by the Australian Research Council (Discovery Project DP0452833) and the Danish Ministry for Science, Technology, and Innovation (VTU) funding for the Centre for Nanostructured Polymer Surfaces for Medical Applications (2002-603-4001-87). LA052143A
