Analysis of Protein Adsorption and Binding at Biosensor Polymer

Electrochemical patterning as a tool for fabricating biomolecule microarrays ... Julia Witt , Britta Vaske , Andreas Lesch , Izabella Brand , Gunther ...
0 downloads 0 Views 395KB Size
Anal. Chem. 2003, 75, 2559-2570

Articles

Analysis of Protein Adsorption and Binding at Biosensor Polymer Interfaces Using X-ray Photon Spectroscopy and Scanning Electrochemical Microscopy Andrew Glidle,† Tomoyuki Yasukawa,† Charlotte S. Hadyoon,† Nathalie Anicet,† Tomokazu Matsue,‡ Masayuki Nomura,§ and Jon M. Cooper*,†

Bioelectronics Research Group, Department of Electronics and Electrical Engineering, University of Glasgow, Glasgow, G12 8LT, U.K., Department of Biomolecular Engineering, Graduate School of Engineering, Tohoku University, Aramaki07, Aoba, Sendai 980-8579, Japan, and Central Technology Laboratory, Asahi Kasei Corporation, 2-1, Samejima, Fuji, Shizuoka 416-8501, Japan

We describe a method, based on X-ray photoelectron spectroscopy (XPS) measurements, to assess the extent of protein adsorption or binding on a variety of different µTAS and biosensor interfaces. Underpinning this method is the labeling of protein molecules with either iodine- or bromine-containing motifs by using protocols previously developed for radiotracer studies. Using this method, we have examined the adsorption and binding properties of a variety of modified electrodeposited polymer interfaces as well as other materials used in µTAS device fabrication. Using polymer interfaces modified with poly(propylene glycol) (PPG) chains, our results indicate that a chain of at least ∼30 monomer units is required to inhibit nonspecific adsorption from concentrated protein solutions. The XPS methodology was also used to probe specific binding of avidins and enzyme conjugates thereof to biotinylated and mixed biotin/PPG-modified polymer interfaces. In one example, using competitive binding, it was established that the mode of binding of a peroxidasestreptavidin conjugate to a biotinylated modified polymer interface was primarily via the streptavidin moiety (as opposed to nonspecific binding via the enzyme conjugate). XPS evaluation of nonspecific and specific peroxidasestreptavidin immobilization on various functionalized polymers has guided the design and fabrication of functionalized interdigitated electrodes in a biosensing µTAS device. Subsequent characterization of this device using scanning electrochemical microscopy (SECM) corroborated the adsorption and binding previously inferred from XPS measurements on macroscale electrodes. Within the field of enzyme-based biosensors and µTAS devices, evaluation of the relative extent of both specific and nonspecific †

University of Glasgow. Tohoku University. § Asahi Kasei Corporation. ‡

10.1021/ac0261653 CCC: $25.00 Published on Web 05/03/2003

© 2003 American Chemical Society

binding of proteins, at both sensor surfaces and on other materials used as system components therein (e.g., microfluidic channels and packaging) is important when interpreting assay results. The most common methods for detecting proteins adsorbed onto such surfaces are optical techniques and include, for example, labeling with fluorescent probes, staining reactions, surface plasmon resonance techniques, and enzyme specific colorimetric assays.1-4 In other work, quantitative assays have also been performed by radio-isotope labeling of protein species.5,6 All of these methods have their relative merits and drawbacks when applied to particular systems. For example, when using fluorescence methods in surface adsorption studies, a prime requirement is that the underlying substrate material has a low background fluorescence. This is generally the case for glasses and colorless polymers but is not always the case for strongly colored materials (e.g., conducting polymers used in biosensors). Likewise, more general protein detection techniques, such as staining reactions, can be compromised by side reactions with functional groups in the polymer substrate that are chemically similar to those in the protein. For example, the Ninhydrin test does not distinguish between the amines in a protein and those on a polymer electrode specifically functionalized with cystamine (via the S terminus) or hydrazide groups (used to immobilize carbohydrate species). An additional hindrance to using colorimetric methods to probe adsorption on some polymer surfaces is the intrinsic color associated with the polymer material. The surface plasmon resonance technique offers high precision in assessing the amount of adsorbed material (on specifically (1) Allen, R. C.; Budowle, B. Protein Staining and Identification Techniques; Eaton Publishing: Natick, MA, 1999. (2) Kessler, C., Ed. Nonradioactive Labeling and Detection of Biomolecules; Springer-Verlag: New York, 1992. (3) Hermanson, G. T. Bioconjugate Techniques; Academic Press: London, 1996. (4) Homola, J.; Yee, S. S.; Gauglitz, G. Sens. Actuators, B 1999, 54, 3-15. (5) Sternberg, R.; Bindra, D. S.; Wilson, G. S.; Thevenot, D. R. Anal. Chem. 1988, 60, 2781-2786. (6) Halliwell, C. M.; Simon, E.; Toh, C.-H.; Bartlett, P. N.; Cass, A. E. G. Biosens. Bioelectron. 2002, 17, 965-972.

Analytical Chemistry, Vol. 75, No. 11, June 1, 2003 2559

modified gold or silver surfaces), but does not directly give information about the type of molecular species adsorbed. It is also found that, because of attenuation of the evanescent wave, the technique cannot be readily used with either strongly colored or thick-film polymer surfaces, such as modified conducting polymer electrodes often used in biosensing or other thick films associated with µTAS or lab-on-a-chip (LOAC) technologies. Enzyme-specific colorimetric assays, similar to those associated with ELISA assays, can be configured to avoid complications associated with the high optical adsorption of a substrate;7 however, such methods detect only the presence of active immobilized biomolecules (and those for which a colorimetric assay exists). Although radio-labeling of biomaterials circumvents many of the problems associated with the optical measurements described above, its application in routine analysis suffers from a number of safety and regulatory issues. Furthermore, the high costs limit the quantity of radio-labeled material available, and the decay lifetime of the label means that material cannot be stored and that measurements must be performed within a short time of sample preparation. As a consequence of the limitations of the techniques described above, we have sought to develop a new method to investigate the nature and extent of protein adsorption or immobilization on modified conducting polymer surfaces. The overall motivation of this study is to develop strategies that could be used to evaluate specific and nonspecific adsorption on the materials used in biosensor or µTAS systems. Previously, X-ray photoelectron spectroscopy (XPS) techniques have been widely used to probe a variety of surfaces, including those of polymers8,9 and monolayers of biomolecules.10-13 However, the technique has been less widely applied to the study of biomolecules on organic polymer surfaces. A probable reason for this is the general similarity between the C (1s) spectra for biomolecules and those of an underlying polymer (for example, in previous investigations of an enzyme-polymer composite, it was necessary to perform a spectral deconvolution in order to assay the system14). We now seek to overcome these difficulties by labeling proteins with appropriate motifs that act as unique spectroscopic markers, such that XPS can be used to probe their presence or absence. In seeking a means to uniquely identify the presence of a protein species in an XPS spectrum, we have modified proteins with either iodine- or bromine-containing labels. These elements have high XPS sensitivities and are not commonly found in either the natural forms of the protein or other materials used in (7) Crowther, J. R., Ed. ELISA: Theory and Practice: Methods in Molecular Biology; Humana Press: Totowa, 1995; Vol 42. (8) Chilkoti, A.; Ratner, B. D. In Surface Characterization of Advanced Polymers; Sabbatini, L., Zambonin, P. G., Eds.; VCH: Weinheim, 1993. (9) Malitesta, C.; Losito, I.; Sabbatini, L.; Zambonin, P. G. J. Electron. Sprectrosc. 1995, 76, 629-634. (10) Pradier, C. M.; Salmain, M.; Zheng, L.; Jaouen, G. Surf. Sci. 2002, 502, 193-202. (11) Facci, P.; Alliata, D.; Andolfi, L.; Schnyder, B.; Kotz, R. Surf. Sci. 2002, 504 (1-3), 282-292. (12) Denis, F. A.; Hanarp, P.; Sutherland, D. S.; Gold, J.; Mustin, C.; Rouxhet, P. G.; Dufrene, Y. F. Langmuir 2002, 18 (3), 819-828. (13) Kingshott, P.; Thissen, H.; Griesser, H. J. Biomaterials 2002, 23 (9), 20432056. (14) Griffith, A.; Glidle, A.; Beamson, G.; Cooper, J. M. J. Phys. Chem. 1997, 101, 2092-2100.

2560

Analytical Chemistry, Vol. 75, No. 11, June 1, 2003

biosensing or µTAS devices. Of the principal labeling methods widely used for labeling proteins with halogens,15-17 for simplicity in this paper, we describe results obtained using Bolton-Hunterbased reagents to react with free amine groups within the protein structure. The basic procedure involves using presynthesized iodine- and bromine-functionalized Bolton-Hunter reagents to label proteins with either Br or I centers, followed by immersing the surface to be examined in the protein solution for a specific period. After rinsing and drying, protein adsorption was evaluated by acquiring XPS spectra in the Br (3d) or I (3d) regions. A variety of materials were investigated, including thin film conducting polymer substrates either in the “as-deposited state”, or after modification with either poly(propylene glycol) (PPG) chains (to inhibit nonspecific adsorption) or biotin (to promote specific avidin-based immobilization). In addition, XPS measurements were made on other surfaces associated with the internal microfluidic components incorporated in the fabrication of µTAS devices (e.g., glasses, polymer resists, siloxane, and silicone-based coverings), providing important information that should be considered in the design of LOAC devices used in proteomics or diagnostics. This XPS method, when calibrated using measurements of adsorption on quartz crystal microbalance (QCM) crystals, was used to provide a quantitative assessment of the amount of protein adsorbed on the surfaces being investigated. The calibration protocol involves measuring the number of XPS counts associated with the halogen atoms of labeled protein that has nonspecifically adsorbed onto the electrode surface of a QCM crystal. This number of counts is then associated with the mass of protein obtained from the shift in frequency of the QCM. The ability of the methodology described here to probe a surface that may contain both Br- and I-labeled species was also used to examine the competitive immobilization of peroxidase and avidin-peroxidase conjugated species on biotinylated surfaces. This information, together with that of the adsorption properties of PPG polymer interfaces was used to design a lithographically fabricated array of conducting polymer-modified electrodes for use within a µTAS device. In this device, avidin-horseradish peroxidase conjugate was immobilized on biotin-modified electrodes within an electrode array- and nonspecific adsorption was prevented from occurring on the other electrodes in the array by modifying with PPG. The redox activity of the peroxidase means that it was possible to use scanning electrochemical microscopy to probe both the microelectrodes and the nearby substrate materials. This last example provided a means by which we could examine whether measurements inferred from the XPS assaying of macroscopic electrodes could be applied to microfabricated devices. MATERIALS AND METHODS Materials. Bis-(2-aminopropyl)propylene glycol 130, 500, 800, and 1900 (NH2PPG130, 500, 800, and 1900) were supplied by Fluka; 5-pentylamine biotin (NH2Biotin) was supplied by Amphotec, Inc. (USA); PEO iodoacetylbiotin, iodogen and dithiobis(succinimidylpropionate) (DSP) were supplied by Perbio; avidin-D (15) Fraker, P. J.; Speck, J. C. Biochem. Biophys. Res. Commun. 1978, 80 (4), 849-857. (16) Bolton, A. E.; Hunter, W. M. Biochem. J. 1973, 133, 529-539. (17) Wilbur, D. S. Bioconjugate Chem. 1992, 3, 433-470.

and avidin-HRP conjugate were supplied by Vector laboratories; horseradish peroxidase (type 1) (HRP), avidin (Av), streptavidin (Strep) and streptavidin-HRP (Strep-HRP) conjugate were supplied by Sigma; alkaline phosphatase (AlkP) was supplied by Roche (U.K.). All other chemicals and reagents used in the procedures described below were obtained from Sigma-Aldrich. Preparation of Diiodo and Dibromo Bolton-Hunter Reagents. 3,5-Dibromo and diiodo derivatives of 4-hydroxyphenylpropionic acid were prepared as described in the literature.18,19 The N-hydroxysuccinimide derivative of the halogenated acid was made using 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride (EDAC)-mediated coupling in DMSO followed by purification using silica gel chromatography (eluting with diethyl ether and 40/60 petroleum ether mixtures). Typically, a 20% excess of N-hydroxysuccinimide was used. Products were characterized by infrared, UV spectroscopy, and elemental analysis using XPS (of a spun thin film on a clean gold substrate) (data not shown). Protein Labeling. The general protocol employed to modify the proteins discussed in the text with iodine- or bromine-labeled Bolton-Hunter reagents was as described in the literature.20 Typically, 1 mg/mL protein solutions were reacted with 0.2 meq/ mg protein of the Bolton-Hunter reagent dissolved in ethyl acetate/propan-2-ol (2:3 v/v). Alternative methods of halogenation (results not shown) included using Iodogen15, chloramine-T (for both iodine17 and bromine labeling21) or direct bromination by adding a 1 mM aqueous solution of NaBr and 1 mM Nbromosuccinimide to the protein solution (in this latter case, 1 mL of bromide solution was added to 1 mL of 1 mg/mL protein solution). In all cases, a reaction time of ∼1 h was used, and excess labeling reagents were removed by exhaustive dialysis using Pierce 10 000 MW dialysis cassettes or gel permeation chromatography using a PD10 column (BioRad). Control experiments using protein-free solutions showed that no residual halogenated species remained after the desalting procedures. Derivatized avidin and streptavidin species were also purified by binding to an iminobiotin agarose column (Sigma) followed by elution with 0.1 M acetic acid. Dialysis was used to exchange the acetic acid for a chosen buffer. Derivatized proteins were assayed as follows: for the modified forms of avidin, streptavidin or the enzyme conjugates, 100-µL aliquots were treated with 5 µL of 0.3 mM 4-hydroxyazobenzene2-carboxiylic acid (HABA). The absorbance at 500 nm was compared to that of solutions of the appropriate underivatized species. Addition of 5 µL of 1 mM biotin solution to these solutions led to the loss of the red color. Horseradish peroxidase was assayed using tetramethylbenzidine (TMB) following the supplier’s protocol. Alkaline phosphatase was assayed using pnitrophenyl phosphate (pNPP), again following the supplier’s protocol. In all cases, the activity of the labeled (and purified) solutions was at least 80% of the unlabeled solutions. Preparation of Modified Polymer Films. 3-(Pyrrol-1-yl)propionic acid (PyCOOH) and its pentafluorophenyl ester deriva(18) Corey, E. J.; Haefele, L. F. J. Am. Chem. Soc. 1959, 81, 2225-2228. (19) Barnes, J. H.; Borrows, E. T.; Elks, J.; Hems, B. A.; Long, A. G. J. Chem. Soc. 1950, 2824-2833. (20) Mock, D. M.; Dubois, D. B. Anal. Biochem. 1986, 153, 272-278. (21) Hadi, U. A. M.; Malcolme-Lawes, J.; Oldham, G. Int. J. Appl. Radiat. Isot. 1979, 30, 709-712.

Scheme 1. Reaction Scheme for Poly(PFP) with Amine-Functionalized Motifs and Subsequent Immobilization of Streptavidin Conjugated Enzymes

tive (PFP) were synthesized using an established method.22 Similarly, the protocol to fabricate poly(PFP) polymer films (∼100 nm thick) by electrochemical oxidation of PFP monomer was as described elsewhere.23 Modification of poly(PFP) films using amine-functionalized species was performed by immersing the polymer electrodes into a 10 mM solution of the appropriate amine dissolved in dimethyl sulfoxide for a period of up to 4 h (Scheme 1). XPS Assaying. XPS measurements were performed using the Scienta ESCA 300 system at RUSTI, Daresbury. General operating conditions, unless otherwise stated, involved using an electron takeoff angle of 90°, analyzer slit width 0.8 mm, energy channel width 0.05 eV, and monochromated Al ΚR radiation. When comparing the magnitudes of XPS signals, care was taken to ensure that each sample was aligned in the same manner in the spectrometer and that there was a stable X-ray flux. By using an electron takeoff angle of 90°, it was possible to ensure that both the maximum depth into the sample was being probed and that small variations in sample angle (up to 5°) had a minimal effect. Integration times were typically 10-20 min for each of the regions C (1s), N (1s), and O (1s) and 20-40 min for the I (3d) or Br (3d) regions. These integration times correspond to times of 1-4 s/0.05 eV channel. For ease of comparison, in the spectra shown, the number of counts collected has been normalized by the collection time and are presented after subtraction of background counts estimated from the signal in the energy channels immediately adjacent to the peak window. On occasion, either to collect better statistics or to investigate the X-ray stability of the labeled protein interface, integration times were increased to a period of a few hours (see Results and Discussion, below). The samples were ∼1 × 1 cm and comprised the material under study deposited or adsorbed onto thin 100-nm (evaporated) gold films, supported on a glass microscope slide substrate. The insulating (22) Pickett, C. J.; Ryder, K. S. J. Chem. Soc., Dalton Trans. 1994, 14, 21812189. (23) Glidle, A.; Bailey, L.; Hadyoon, C. S.; Hillman, A. R.; Jackson, A.; Ryder, K. S.; Saville, P. M.; Swann, M. J.; Webster, J. R. P.; Wilson, R. W.; Cooper, J. M. Anal. Chem. 2001, 73 (22), 5596-5606.

Analytical Chemistry, Vol. 75, No. 11, June 1, 2003

2561

Scheme 2. Reaction of Poly(PFP) with Cystamine, Followed by Coupling of PEO-Iodoacetyl Biotin

nature of the substrate necessitated the use of a flood gun (set to 2 eV), and consequently, the presented spectra have been corrected by assigning 285 eV to the main C-H peak in the C (1s) spectra. Quantification of Absolute Surface Coverages. Quartz crystal microbalance (QCM) measurements were performed using the instrumentation and procedures detailed elsewhere.24 Goldcoated QCM crystals (10 MHz AT cut, polished) were supplied by ICM, Oklahoma City. Prior to measurements, an Ar+ etch pretreatment was used to remove all traces of contamination from the surface of the gold-coated QCM. In a typical experiment to determine the conversion factor between the number of XPS counts in the I (3d) (or Br (3d)) region and the protein surface coverage, XPS and QCM resonant frequency measurements were performed on dry crystals both before and after immersion in 1 mg/mL of labeled protein for ∼1 h. Thorough washing in pH 7.2 phosphate buffered saline (PBS), followed by rinsing in water served to both remove electrolyte and ensure that multilayer films of protein were not formed on drying. To overcome complications that may arise in the interpretation of crystal frequency changes as being adsorbed mass (e.g., due to the presence of water in the adsorbed layer and possible viscoelastic effects in solution), all crystal frequency measurements were made in an evacuated desiccator (10-3 Torr). Measurements were made to determine the AC impedance of the QCM crystal in the vicinity of the resonance frequency both before and after protein adsorption using the procedure described elsewhere.24 The lack of change in the peak shape of the impedance response indicates that the adsorbed protein can be considered as a rigid layer on the surface of the crystal (see Supporting Information Figure 1). Nonspecific Binding of Proteins on Modified Polymer and Gold Surfaces. A series of poly(PFP) films were modified with PPG adducts of different chain lengths (Scheme 1). The modified films and as-deposited polymers were immersed in pH 7.2 solutions of iodinated or brominated protein for predetermined periods of time. The samples were thoroughly washed in 0.1 M PBS, pH 7.2, followed by water and were blow-dried using N2 immediately prior to XPS measurements. Modification of bare gold electrode surfaces with PPG chains was performed in two stages. First, a N-hydroxysuccinimide functionalized self-assembled monolayer was formed on clean (Ar etched) evaporated gold surfaces by immersion for 2 h in a solution of 10 mM dithiobis(succinimidylpropionate) (DSP) in dimethylformamide (DMF). After rinsing with DMF and drying with N2, the substrates were immersed for 3 h in 10 mM solutions of the desired NH2-terminated PPG dissolved in DMSO. The (24) Glidle, A.; Hadyoon, C. S.; Cass, A. E. G.; Cooper, J. M. Electrochim. Acta 2000, 45, 3823-3831.

2562

Analytical Chemistry, Vol. 75, No. 11, June 1, 2003

substrates were rinsed with DMSO and dried with N2 prior to use in nonspecific binding studies. Concentration and Time-Dependence of Protein Adsorption. Modified polymer interfaces were immersed in different ionic strength solutions (0.01-1 M) containing various concentrations (10 µg/mL - 1 mg/mL) of the iodinated or brominated protein under investigation for specified periods of time (10 min - 3 h). XPS measurements were made following exhaustive washing and drying, as above. Protein Adsorption on µTAS Materials. The NSB of a variety of proteins (including avidin, streptavidin, streptavidinHRP and alkaline phosphatase) on a series of different materials used in the fabrication biosensor or µTAS devices was determined using XPS measurements, as detailed above. The materials examined included glass, silicone rubber (Dow Corning 3140 and 3145), poly(dimethylsiloxane) (Dow Corning, Sylgard Elastomer 184), and the thick film photoresist SU-8 (MicroChem Corp., USA). The preparation and pretreatment of these various materials was the same as when they are used in microfabrication protocols. Specific Binding to Biotinylated Surfaces. Poly(PFP) surfaces were derivatized with pentylamine biotin using a protocol similar to that for derivatization with PPG chains (Scheme 1). The derivatized films were then immersed in iodinated or brominated avidin or streptavidin solutions, as described in the text, in order to study the extent of specific or “directed” immobilization at polymer interfaces. To investigate the effect on avidin or streptavidin binding of increasing the length and hydrophilicity of the spacer arm used (in coupling the biotin motif to polymer surface), a two step functionalization procedure was used. First, the poly(PFP) film was reacted with cystamine to produce a surface with potentially free thiol motifs (Scheme 2). This was then further reacted with PEO-iodoacetyl biotin to generate the biotinylated surface. It was found that the iodoacetyl reaction was most efficient if the cystamine-functionalized polymer surface was pretreated with deoxygenated dithiothretol (DTT) and then thoroughly washed with deoxygenated water prior to immersion in the iodoacetyl solution (this procedure was performed in a nitrogen filled glovebag). Reflectance FT-IR and XPS measurements (C, N, O, and S regions) were performed at each stage to confirm that the expected reaction had occurred. Mixed Functionalized Layers of Biotin and PPG. To make an estimate of the extent of specific and nonspecific binding on biotinylated surfaces, XPS measurements were made following immersion in either iodine-labeled streptavidin (I-strep) solutions or I-strep solutions containing 5 mM of dissolved biotin (designed to block molecular recognition sites on solution based streptavidin). In an alternative approach, attempts were made to displace

Scheme 3. Characterization of HRP-Avidin Immobilized on Functionalized Interdigitated Array Electrodes Using Scanning Electrochemical Microscopy (SECM)a

a

Blue electrodes correspond to biotin-functionalization of poly(PFP); red electrodes are PPG-modified poly(PFP) (see Scheme 1). Microelectrode diameter, 9 µm; scan rate, 20 µm/s; microelectrode potential, -0.05 V vs Ag|AgCl. Scanned area, 800 × 300 µm; spatial resolution, 10 µm. Solution, 2.0 mM FcOH, 1.0 mM H2O2.

specifically adsorbed I-avidin or I-streptavidin by soaking biotinylated films in 10 mM biotin solutions for ∼3 h after exposure to I-avidin or I-streptavidin solution. To further investigate the specificity of avidin binding, surfaces were prepared containing a mixture of PPG and biotin functionalities with a moderate to high proportion of biotin motifs by immersing poly(PFP) films in solutions containing both NH2PPGs and NH2Biotin (in molar ratios from 1:1 to 50:1). Dilute biotinylated surfaces were prepared by reacting EDAC-activated biotin with PPG1900-functionalized poly(PFP) films for 12 h. In both cases, the formation of a surface with mixed functionality was confirmed by XPS and reflectance FT-IR. Competitive Adsorption of Streptavidin and StreptavidinHRP. Biotin-modified electrodes were immersed in solutions of either bromine-labeled HRP, iodine-labeled streptavidin-HRP conjugate or 50:50 mixtures of bromine-labeled HRP and iodinelabeled streptavidin-HRP conjugate. Total protein concentrations of either 0.01 or 0.1 mg/mL in 0.1 M PBS, pH 7.2, were used, together with an immersion time of 100 min. XPS measurements were then performed after rinsing and drying as above. SECM Measurements. Two sets of interdigitated 100-µmwide band electrodes were fabricated on a glass substrate using standard photolithographic methods.25 Poly(PFP) was deposited onto one set of electrodes,23 which was then functionalized with NH2Biotin. Following complete functionalization, poly(PFP) was deposited onto the second set of electrodes, and the substrate was immersed in NH2PPG1900. This resulted in an alternating sequence of band electrodes that were alternately functionalized by biotin and PPG1900. The entire functionalized interdigitated substrate was then immersed in a solution of 0.1 mg/mL avidinHRP conjugate for a period of ∼100 min. To investigate whether “active” HRP immobilization was confined to the biotinylated microelectrodes, the SECM experiment depicted in Scheme 3 was performed. In this experiment, the substrate was immersed in a 2 mM ferrocenemethanol (FcOH) solution containing 1 mM hydrogen peroxide. The local concentration of electroactive species was measured by scanning the microelectrode over the sample surface. The instrumentation used was as described previously.26 (25) Rai-Choudhury, P., Ed. Handbook of Microlithography, Micromachining, and Microfabrication; Volume 2: Micromachining and Microfabrication; SPIE: CITY OF PUBLICATION, 1999.

FT-IR Measurements. Reflectance adsorption infrared (RAIRS) measurements of large samples (∼1 cm2) were made using a Bomem MB120 instrument equipped with a variable angle specular reflectance accessory (set to 74°). p-Polarized incident radiation was selected using a ZnS polarizer (Grasbey-Specac). Reflectance microscope measurements were performed using a Nicolet system comprising a continuum IR microscope with cooled MCT detector, working in specular reflectance mode, attached to a Nexus research spectrometer. RESULTS AND DISCUSSION In the majority of the studies described below, we have exploited a generic method that enabled us to create a conducting polymer surface which can be subsequently modified with a specialized structural motif. The general protocol is shown in Scheme 1. One attractive feature of this strategy is that only one electropolymerizable monomer need be synthesized, pentafluorophenyl 3-(pyrrol-1-yl)propionate (PFP), to produce a variety of interfacial assemblies. A second advantage of using this system is that it provides a method whereby electrodes within microstructures can be selectively modified with different surface adducts. Probing Nonspecific Binding on Modified Polymers Using XPS. Previously, both avidin and streptavidin have been used to create controlled interfacial assemblies at either biotinylated gold or polymeric surfaces.27 Such architectures, when incorporating conjugated HRP, have formed the basis of ELISA measurements. Important in such measurements is the inhibition of nonspecific binding, and both poly(propylene glycol) and poly(ethylene glycol) have been proposed as suitable adducts to overcome protein nonspecific binding at interfacial assemblies.13,28,29 Thus, as part of the process of characterizing the surface properties of materials proposed for use as immobilization and detection electrodes in a biosensing µTAS device based on the specific binding of avidinHRP conjugate to biotinylated conducting polymers, we first describe the simple comparison of the extent of the nonspecific binding of avidin and HRP onto poly(propylene glycol)-modified polymers with that found on as-deposited conducting polymer surfaces. One overall aim of the work was to create a modified conducting polymer surface on which there was both a greatly reduced level of nonspecific adsorption of protein and the conducting region of the polymer electrode was still accessible to small solution species for electron transfer purposes. We, therefore, investigated polymer surfaces modified with amineterminated poly(propylene glycol) (PPG) chains having a variety of different chain lengths. The terminal amine groups on the PPG adduct act as nucleophiles in the bimolecular reaction with the activated ester contained within poly(PFP) (Scheme 1). In this study, we used a series of readily available amine-terminated PPG chains that have average molecular weights between 130 (2 monomer units) and 1900 (32 monomer units). (26) Yasukawa, T.; Kaya, T.; Matsue, T. Anal. Chem. 1999, 71, 4637-4641. (27) Savage, M. D.; Mattson, G.; Desai, S.; Nielander, G. W.; Morgensen, S.; Conklin, E. J. Avidin-Biotin Chemistry: A Handbook; Pierce Chemical Company: CITY OF PUBLICATION, 1992. (28) Otsuka, H.; Nagasaki, Y.; Kataoka, K. Curr. Opin. Colloid 2001, 6 (1), 3-10. (29) Lu, H. B.; Campbell, C. T.; Castner, D. G. Langmuir 2000, 16 (4), 17111718.

Analytical Chemistry, Vol. 75, No. 11, June 1, 2003

2563

Figure 1. (a) I (3d) XPS spectra for a series of polymer electrodes after immersion for 100 min in 0.1 mg/mL iodo-avidin solution (pH 8.5, 0.1 M phosphate buffer). Spectra of poly(pyrrole) and poly(PFP) (marked Py and PFP) are nearly coincident. I (3d) signal of PPG 800modified electrode is slightly larger than that of the PPG 1900 electrode. (b) Br (3d) XPS spectra for a series of polymer electrodes after immersion for 100 min in 0.1 mg/mL bromo-HRP solution (pH 7, 0.1 M phosphate buffer).

The I (3d) XPS spectra of Figure 1a show that avidin adsorption was largest on films of as-deposited poly(PFP), poly(pyrrole), and poly(PFP) modified with the short-chain PPG130. For this particular protein and solution conditions, it can be seen that modification with PPG1900 is required to reduce the NSB to almost negligible levels. Close inspection of the XPS spectra of the adsorbed labeled protein reveals small shoulders on the low binding energy side of the 621 eV/632.5 eV I (3d) doublet, which may be attributable to I- species. Control experiments using KI electrolyte in the absence of protein show that this spectroscopic feature does not arise from I- ions adsorbed on to the polymer from the electrolyte. It is possible that these species may result from a small amount of X-ray-induced cleavage of C-I bonds. Importantly for the quantification presented below, repetitive recording of spectra in this region did not result in a significant decrease in the overall numbers of I (3d) counts or significant further decomposition. Figure 1b shows a similar set of spectra recorded in the Br (3d) doublet region for bromine-labeled HRP adsorbed onto various conducting polymer substrates. These spectra show that the presence of brominated species is clearly measurable, and 2564

Analytical Chemistry, Vol. 75, No. 11, June 1, 2003

again, modification of the surface with intermediate length PPG chain reduces the amount of HRP adsorbed. Quantification of Absolute Surface Coverages. In performing the above measurements, it was possible to estimate the number of atomic centers probed in a particular sample by summing the number of electron counts measured in the C (1s), N (1s), O (1s), and I (3d) or Br (3d) regions (after proportioning using the sensitivity factors associated with the different electron shells). This analysis showed that the variation in the total number of atomic centers sampled (excluding hydrogen) was within (2% for a set of samples of different surface functionality but of similar elemental composition, measured in a particular series of experiments (the variation is (1.8% if elements are weighted for electron adsorption). The consistency in numbers of atoms probed is a consequence of a number of factors: (i) all samples were of similar (electron) density, and so the mean free paths are similar (both the polymer and proteins are predominantly composed of C, H, N, and O); (ii) samples had a similar surface morphology; (iii) in modern XPS instrumentation, both the X-ray source and the detector, as well as the mechanical mounts, are significantly more stable than in previous generations of instruments; and (iv) all measurements in a particular experimental series were performed within 1 day. The last two of these factors are corroborated by our findings that when particular samples were measured, removed from the instrument, and subsequently remeasured, the numbers of counts varied by