Dynamic, Electronically Switchable Surfaces for Membrane Protein

We propose a novel approach to create a (membrane) protein microarray by using an indium tin oxide (ITO) microelectrode array with an electronic multi...
0 downloads 0 Views 664KB Size
Anal. Chem. 2006, 78, 711-717

Dynamic, Electronically Switchable Surfaces for Membrane Protein Microarrays C. S. Tang,†,‡ M. Dusseiller,‡ S. Makohliso,§ M. Heuschkel,§ S. Sharma,‡ B. Keller,† and J. Vo 1 ro 1 s*,‡

Swiss Federal Laboratories for Materials Testing and Research (EMPA), Du¨bendorf, Switzerland, BioInterface Group, Laboratory for Surface Science and Technology, Swiss Federal Institute of Technology (ETH), Zurich, Switzerland, and Ayanda Biosystems SA, Lausanne, Switzerland

Microarray technology is a powerful tool that provides a high throughput of bioanalytical information within a single experiment. These miniaturized and parallelized binding assays are highly sensitive and have found widespread popularity especially during the genomic era. However, as drug diagnostics studies are often targeted at membrane proteins, the current arraying technologies are ill-equipped to handle the fragile nature of the protein molecules. In addition, to understand the complex structure and functions of proteins, different strategies to immobilize the probe molecules selectively onto a platform for protein microarray are required. We propose a novel approach to create a (membrane) protein microarray by using an indium tin oxide (ITO) microelectrode array with an electronic multiplexing capability. A polycationic, protein- and vesicle-resistant copolymer, poly(L-lysine)-grafted-poly(ethylene glycol) (PLL-g-PEG), is exposed to and adsorbed uniformly onto the microelectrode array, as a passivating adlayer. An electronic stimulation is then applied onto the individual ITO microelectrodes resulting in the localized release of the polymer thus revealing a bare ITO surface. Different polymer and biological moieties are specifically immobilized onto the activated ITO microelectrodes while the other regions remain protein-resistant as they are unaffected by the induced electrical potential. The desorption process of the PLL-g-PEG is observed to be highly selective, rapid, and reversible without compromising on the integrity and performance of the conductive ITO microelectrodes. As such, we have successfully created a stable and heterogeneous microarray of biomolecules by using selective electronic addressing on ITO microelectrodes. Both pharmaceutical diagnostics and biomedical technology are expected to benefit directly from this unique method. Microarrays are bioanalytical assays consisting of a planar substrate with regular rows and columns of microscopic elements/ spots to immobilize specific molecular probes or cells in solution. The fundamental principle to miniaturize assay methodologies * To whom correspondence should be addressed. Telephone: + 41 1 632 59 03. Fax: + 41 1 633 10 27. E-mail: [email protected]. † Swiss Federal Laboratories for Materials Testing and Research. ‡ Swiss Federal Institute of Technology. § Ayanda Biosystems SA. 10.1021/ac051244a CCC: $33.50 Published on Web 12/17/2005

© 2006 American Chemical Society

could be traced back to more than a decade ago when Ekins proposed the “ambient analyte theory”.1 At that time, it was demonstrated that, within certain limitations, microspot assay signal density increased with decreasing amount of captured molecules, hence enabling it to be more sensitive for various detection techniques, compared to any other ligand-binding assay.2 Since then, microarrays have been widely employed for diverse biomedical research such as gene expression profiling,3 cancer,4 cellular networks,5and proteomics.6 One of the strongest motivations that fueled the growth of the microarray technology is derived from the success of the deoxyribonucleic acid (DNA) biochips. Various approaches were suggested for designing a DNA biosensing platform,7,8 and among them, a novel method proposed by Nanogen involves an electronic control.9 This technique enables selective electronic addressing of a positive bias onto microelectrodes to actively transport and concentrate negatively charged DNA molecules within a well-defined electrically activated microspot. To ensure hybridization stringency a negative bias was applied to remove any unhybridized molecules from the microelectrodes. The human genome project is in its final phases of completion now, and there is an emerging focus toward understanding the functions of proteins. However, several studies demonstrated the poor correlation between the messenger ribonucleic acid and protein expression.10,11 Due to the inherent complex 3-D structure of proteins as well as the absence of any protein amplification tools,12 it is difficult to replicate the lessons learned from the DNA microarray technology and apply them for the proteomics microarray. Some of the problems encountered when dealing with arraying proteins are as follows: current spotting technology of fragile proteins onto a microarray involves excessive shear forces and drying conditions that could denature the stability of the (1) Ekins, R. P. J. Pharm. Biomed. Anal. 1989, 7, 155-168. (2) Templin, M. F.; et al. Trends Biotechnol. 2002, 20, 160-166. (3) Ryan, M. M.; et al. Biol. Psychiatry 2004, 55, 329-336. (4) Kruse, J. J. C. M.; te Poele, J. A. M.; Russell, N. S.; Boersma, L. J.; Stewart, F. A. Int. J. Radiat. Oncol., Biol., Phys 2004, 58, 420-426. (5) Scholl, M.; et al. J/ Neurosci. Methods 2000, 104, 65-75. (6) Tan, L.-P.; Chen, G. Y. J.; Yao, S. Q. Bioorg. Med. Chem. Lett. 2004, 14, 5735-5738. (7) Lian, W.; et al. Anal. Biochem. 2004, 334, 135-144. (8) Srivannavit, O.; et al. Sens. Actuators, A 2004, 116, 150-160. (9) Edman, C.; et al. Nucleic Acids Res. 1997, 25, 4907-4914. (10) Lueking, A.; et al. Anal. Biochem. 1999, 270, 103-111. (11) Gygi, S.; Rochon, Y.; Franza, R.; Aebersold, R. Mol. Cell. Biol. 1999, 19, 1720-1730. (12) Bos, M. A.; et al. Colloids Surf., B: Biointerfaces 1994, 3, 91.

Analytical Chemistry, Vol. 78, No. 3, February 1, 2006 711

molecules and hence results in the loss of its functionality and activity. In addition, studies have shown that proteins are not strongly influenced by external electrical potential, and this reduces the possibility to directly immobilize proteins with electrostatic strategies.12 In the design of a protein microarray, it is also critical to include a chemically well-defined substrate surface in order to ensure a homogeneous and stable ligand assembly onto the interface.13 Conventional protein immobilization techniques are widely available, and they comprise adsorption onto a solid support, entrapment within a membrane, and covalent binding to a support.14 However, contaminant protein adsorption onto surfaces often competes with selective protein-substrate interactions, and this leads to an ambiguous background signal and inactivity of the entrapped protein. To overcome nonspecific protein adsorption, a successful strategy is to render the surrounding area proteinresistant, e.g., with a polycationic copolymer, poly(L-lysine)-graftedpoly(ethylene glycol) (PLL-g-PEG).15,16 By end-functionalizing the PEG chains of the graft copolymer, PLL-g-PEG, with biotinylated molecules, specific bioaffinity sensing can be achieved on a biosensor platform.17 A dynamic feature that can be integrated within the design of a microarray platform is an electrical switching option to modify surface properties for specific functions such as hydrophobicity. Recently, Lahann et al. demonstrated that a low density of selfassembled monolayer established with sufficient spatial freedom on a gold surface was electrically stimulated to undergo conformational transitions between a hydrophilic and moderately hydrophobic state.18 Yeo et al. showed that an electroactive substrate dynamically releases cells and mitigates cellular growth and migration with a low electrical potential applied on the surfacemodified patterned platform.19 Using simple surface chemistry and selective electronic addressing, Kim et al. introduced a singlestep electrochemical covalent coupling method to produce density gradients of streptavidin patterns.20 Wa¨lti et al. showed that gold electrodes could be selectively functionalized on a nanoscale by the electrochemical desorption of a molecular protection layer, followed by the subsequent adsorption of the thiolated oligonucleotides.21 In this paper, we demonstrate a novel strategy to electronically address an indium tin oxide (ITO)-based microelectrode array, which is uniformly coated with protein-resistant polymer, PLL-gPEG. The induced electrical field is confined within the selected microelectrodes, and thus, electrochemically removes the PLLg-PEG adlayer only from its conductive surfaces. The surrounding insulating area is totally unaffected by the electrical potential and remains protein-resistant. Subsequently, the electrically activated (13) Houseman, B. T.; Mrksich, M. Trends Biotechnol. 2002, 20, 279-281. (14) Williams, R. A.; Blanch, H. W. Biosens. Bioelectron. 1994, 9, 159-167. (15) Huang, N.-P.; Michel, R.; Voros, J.; Textor, M.; Hofer, R.; Rossi, A.; Elbert, D. L.; Hubbel, J. A.; Spencer, N. D. Langmuir 2001, 17, 489. (16) Michel, R.; Lussi, J. W.; Csucs, G.; Reviakine, I.; Danuser, G.; Ketterer, B.; Hubbell, J. A.; Textor, M.; Spencer, N. D. Langmuir 2002, 18, 3281-3287. (17) Huang, N. P.; Voros, J.; De Paul, S. M.; Textor, M.; Spencer, N. D. Langmuir 2002, 18, 220-230. (18) Lahann, J. Science 2003, 300, 903-903. (19) Yeo, W. S.; Hodneland, C. D.; Mrksich, M. Chembiochem 2001, 2, 590-+. (20) Kim, K.; Jang, M.; Yang, H.; Kim, E.; Kim, Y. T., Kwak, J. Langmuir 2003, 20, 3821-3823. (21) Walti, C.; Wirtz, R.; Germishuizen, W. A.; Bailery, D. M. D.; Pepper, M.; Middelberg, A. P. J.; Davies, A. G. Langmuir 2003, 19, 981-984.

712

Analytical Chemistry, Vol. 78, No. 3, February 1, 2006

ITO microelectrodes can be surface-functionalized with different biological moieties of interest, such as DNA, proteins, or lipid vesicles containing membrane proteins. We have used fluorescently labeled fibrinogen and streptavidin to show the proof of concept for creating fluorescently labeled protein and vesicle arrays to demonstrate the potential for the creation of membrane protein microarrays. RESULTS Fluorescently labeled protein-resistant copolymer PLL-g-PEG/ 633 was used as a passivating adlayer to prevent nonspecific protein adsorption onto surfaces. The presence of the fluorescent label on the molecule enabled the visualization of the uniformly adsorbed polymer layer on a 40 × 40 µm rectangular, electrically conductive, indium tin oxide electrode and on the surrounding insulating epoxy surfaces, by using confocal laser scanning fluorescence microscopy (Figure 2, inset a). The sidewalls appeared brighter because of their vertical geometry (5 µm), which resulted in a higher amount of fluorescent molecules in the detection volume of the focused laser. Figure 2 depicts the initial fluorescence peak intensity of both the ITO microelectrode and the epoxy background normalized at 100%. When an electrical polarization of + 1800 mV was applied onto the ITO microelectrode, we observed a decay of fluorescence intensity on the ITO spot. It was noted that the steady loss of fluorescence signal followed a single-exponential decay with a time constant of 8.2 s. After 24 s of an applied potential at + 1800 mV, the relative intensity value on the polarized ITO reduced to 10% and became almost indistinguishable from the signal of a photobleached region. This indicated a significant desorption of the PLL-g-PEG/633 from the electrically conductive ITO surface (Figure 2, inset b) and hence exposed the ITO surfaces for subsequent surface functionalization. Only a slight loss of fluorescence intensity was observed on the insulating epoxy region due to the photobleaching of the fluorophores by the laser. After an electrical stimulus, different types of functionalized polymers or self-assembled monolayer could be adsorbed onto the resulting bare ITO microelectrodes. As a demonstration, the microelectrode array (MEA) was exposed to a different labeled polymer, PLL-g-PEG/488, and thoroughly rinsed. Figure 3a shows that the selected set of ITO microelectrodes that were electronically activated earlier were now uniformly adsorbed with fluorescently labeled PLL-g-PEG/488. The inactivated ITO microelectrodes and the silica region were unaffected by the electrical stimulation and remained covered with PLL-g-PEG/633. The ITO microelectrodes, which have been covered with an adlayer of PLL-g-PEG/488, were subjected to + 1800 mV again and it resulted in the removal of the polymer from the ITO surfaces (Figure 3b). We noted that the integrity of the ITO microelectrodes was not compromised, despite subjecting it to repeated electronic activation. Four inactivated microelectrodes on the same platform were also subjected to +1800 mV (white arrows in Figure 3c). We observed that all the electronically activated ITO microelectrodes on the MEA have an intensity signal similar to that of a photobleached region (Figure 3c). For our next experiment, we uniformly coated the MEA surfaces with biotinylated protein-resistant polymer, PLL-g-PEG/ PEGbiotin. An electrical polarization of +1800 mV was applied onto a series of ITO surfaces and the protein-resistant polymer

Figure 1. Schematic view of creating a heterogeneous microarray with dynamic capabilities. (a) A glass substrate with neighboring regions of electrically conductive ITO and insulating epoxy or silicon oxide was fabricated for a MEA. Two possible types of protein-resistant surface could be adopted for the MEA: a solution of fluorescently labeled PLL-g-PEG/633 (b) or biotinylated PLL-g-PEG/PEGbiotin (c). (d) An electrical polarization of +1800 mV (referenced to a silver electrode) is selectively applied onto the individual microelectrode. The protein-resistant polymer on the conductive region of the ITO is removed from its surface without affecting the integrity of the adlayer on the insulating epoxy surface. (e) The MEA is exposed for subsequent biological immobilization. (i) In the first example, a different polymer, PLL-g-PEG/488, was introduced into the MEA and it was found to adsorb specifically onto the previously electrically polarized ITO surface. (ii) For the second experiment, the MEA was subsequently exposed to a solution of fluorescently labeled streptavidin/633, which has a strong binding affinity to the biotinylated vesicles, DOPC/488. (iii) In another study, a solution of fluorescently labeled protein, fibrinogen/488, was introduced into the MEA, which adsorbed only onto the exposed ITO microelectrodes. Then another solution of fluorescently labeled protein, streptavidin/633, was introduced into the MEA, which has a specific binding only onto the regions coated with an adlayer of biotinylated protein-resistant polymer, PLL-g-PEG/PEGbiotin.

was removed from the ITO surface, hence exposing a platform for subsequent biological immobilization. Figure 4a shows that when the MEA was exposed to fluorescently label protein, fibrinogen/488, only the electronically activated microelectrodes displayed green fluorescence intensity while the rest of the MEA remained dark. This indicates that the labeled proteins adsorbed specifically onto the bare ITO microelectrode surfaces while the surrounding region, which was covered with the biotinylated protein-resistant polymer PLL-g-PEG/PEGbiotin, prevented nonspecific protein adsorption. A different fluorescently label protein, streptavidin/633, was later introduced into the flow cell. Since streptavidin has a specific binding affinity to biotin, only the regions on the MEA that were adsorbed with PLL-g-PEG/ PEGbiotin displayed distinctively red fluorescence intensity (Figure 4b). No interaction between the two fluorescently labeled proteins was observed outside the polarized ITO microelectrodes,

due to the presence of an adlayer of protein-resistant polymer. However, it was noted that the perimeter surrounding all the ITO microelectrodes has higher fluorescence intensity. As mentioned earlier, this phenomenon was primarily attributed to the fluorescently labeled protein streptavidin/633, which was adsorbed onto the vertical epoxy walls (Figure 4c). To contribute to the technology for the membrane protein microarray, we conducted a feasibility experiment to explore the selective immobilization of vesicles. The physical adsorption of neutral DOPC vesicles onto an ITO surface resulted in weak and unstable adhesion contacts, hence affecting the structure and activity of the vesicles. As such we adopted an interfacial strategy of using a streptavidin-coated surface to bind biotinylated vesicles onto the ITO microelectrodes. For arraying the lipid vesicles, PLLg-PEG/633 was selectively desorbed from two ITO microelectrodes at +1800 mV (Figure 5a and b). Both the insulating region Analytical Chemistry, Vol. 78, No. 3, February 1, 2006

713

Figure 2. Relative fluorescence intensities on the surface-modified ITO microelectrode and the insulating epoxy region, with respect to time. At rest potential, the fluorescence signal on both the ITO microelectrode and epoxy region was normalized at 100%. When an electrical polarization was applied onto the ITO microelectrode (at ∼10 s), the fluorescence intensity exponentially decreased to 10% after 24 s. This indicated a rapid desorption of the fluorescently labeled protein-resistant polymer PLL-g-PEG/633 specifically from the polarized ITO microelectrode. Insignificant loss of fluorescence intensity due to photobleaching was observed on the insulating epoxy region, which inferred that it remained protein-resistant. Inset: Confocal fluorescence images of the following: (a) A 40 × 40 µm ITO microelectrode surrounded by insulating epoxy on a MEA. The surface of the MEA was exposed to fluorescently labeled protein-resistant copolymer PLL-g-PEG/633 for 30 min at ambient conditions and rinsed. Uniform fluorescence intensity was observed on the ITO MEA. A stronger fluorescence signal was present on the perimeter of the microelectrode due to their vertical geometry (5 µm), which resulted in a higher amount of fluorescent molecules in the detection volume of the focused laser. (b) An electrical stimulus of +1800 mV (reference to a silver electrode) on the ITO microelectrode removed the PLLg-PEG/633 after 24 s. No loss of fluorescence intensity was observed on the surrounding area of the ITO microelectrode.

and inactivated microelectrodes remained protein-resistant. When fluorescently labeled streptavidin/633 was injected into the flow cell, we observed that the proteins adsorbed specifically onto the two selected ITO microelectrodes that have been electronically activated (Figure 5c). Subsequently, biotinylated fluorescently labeled DOPC/488 vesicles were introduced and they were found to bind specifically onto the two ITO microelectrodes that were covered with an adlayer of streptavidin/633 (Figure 5d). DISCUSSION ITO is commonly used in the optoelectronic field, primarily due to its unique properties of high electrical conductivity and optical transparency. In our case, the use of ITO as an electronic platform for dynamic biosensing applications provides several direct advantages. First, ITO MEAs and the interface units that allow for electronic addressing of each microelectrode on a selective basis are commercially available. Second it enables in situ imaging techniques such as confocal laser scanning microscopy to investigate the effects of an external electrical stimulus on a surface-modified ITO MEA. 714

Analytical Chemistry, Vol. 78, No. 3, February 1, 2006

Based on our experiments, we have demonstrated that a surface-modified ITO microelectrode could be activated with an external electrical potential for subsequent immobilization of proteins or vesicles. The integrity of the ITO microelectrodes was not compromised despite several repeated electronic activations. This reversible switching process has the potential for multiple analyte screening and could significantly increase the throughput of proteomic analysis. Currently, we are not aware of any similar method that employs an external electrical polarization as an electronic switching means to reversibly remove a passive adlayer for biological surface functionalization. The confocal fluorescence microscopy images revealed a distinctive resolution between the fluorescence intensities on the 40-µm ITO microelectrodes and the surrounding region. Previous work done in our laboratory has shown that it is possible to achieve a high image resolution for electrochemical patterning with fidelity below 2 µm resolution. This is indicative that the electrical polarization is highly localized and confined within the conductive spot, without affecting the neighboring insulating region. However, we believe that the resolution for electrochemical patterning is not limited to the micrometer range and it could be further reduced down to the nanometer scale for single-molecule or single-vesicle immobilization. At this stage, we are unclear on the exact type of desorption mechanisms that could account for the electrochemical removal of the protein-resistant polymer from the ITO microelectrode surface when an external electrical potential of +1800 mV was applied. Ongoing studies are currently in progress to further elucidate the desorption mechanism of the polymer from the electrically polarized ITO surface. Nevertheless, we attempt to postulate two possible explanations for the electrochemical desorption of the polycationic polymer PLL-g-PEG from the polarized ITO surface: accumulation of surfaces charges and reactive oxygen evolution. When a positive electrical polarization is applied onto the ITO, a buildup of positive surface charges accumulates at the solid region solid/liquid interface. With +1800 mV, the strength of the excessive electrical charges extends beyond the capacitive double layer of negative ions and interacts directly with the positively charged backbone chain of PLL. Hence, this discourages the stable adhesion of the protein-resistant polymer PLL-g-PEG and desorbs the polymer from the electrically conductive ITO surfaces. On the other hand, with an electrical polarization at +1800 mV (reference to silver electrode), water dissociation effects are known to occur. We postulate that, during this process, oxygen evolution from the ITO surface takes place and dissolved reactive oxygen could introduce chemical changes weakening the electrostatic interactions between PLL-g-PEG and the ITO surface. Hence, this would release the polymer from the ITO microelectrode. In addition, during the oxygen evolution, the presence of positively charged hydrogen ions accumulating on the ITO surface could result in a localized drop in pH on the ITO surface and this would then repel the polycationic polymer PLL-g-PEG. MATERIALS AND METHODS Materials. The MEA was fabricated accordingly as previously reported.22 It was composed of a glass substrate with an 8 × 8 (22) Heuschkel, M. O.; Fejtl, M.; Raggenbass, M.; Bertrand, D.; Renaud, P. J. Neurosci. Methods 2002, 114, 135-148.

Figure 3. Confocal fluorescence images. (a) A MEA uniformly adsorbed with fluorescently labeled protein-resistant polymer PLL-g-PEG/633. A series of ITO microelectrodes were then subjected to an electrical stimulation, and the polymer was removed from the selected ITO microelectrodes. The MEA was subsequently exposed to another fluorescently labeled protein-resistant polymer, PLL-g-PEG/488, and rinsed. The polymer, PLL-g-PEG/488 (green), was observed to adsorbed specifically only onto the electrically polarized ITO microelectrodes while the surrounding region remained covered with PLL-g-PEG/633 (red). (b) The ITO microelectrodes with PLL-g-PEG/488 (green) were electrically stimulated again, and this removed the polymer from the surfaces, hence revealing the underlying background signal. (c) Subsequently, four inactivated ITO microelectrodes (white arrows) were electrically polarized to remove the adlayer of PLL-g-PEG/633 (red).

Figure 4. Confocal fluorescence images. (a) The MEA was modified with biotinylated PLL-g-PEG/PEGbiotin. The bare surfaces of the ITO microelectrodes were exposed after an electrical polarization of +1800 mV. When the MEA was exposed to fluorescently label protein fibrinogen/ 488, only the electronically activated regions of the ITO microelectrodes displayed green fluorescence intensity while the rest of the MEA remained dark. (b) Another fluorescently labeled protein solution, streptavidin/633, was introduced into the MEA. Since streptavidin has a specific binding affinity with biotin, only the regions on the MEA adsorbed with PLL-g-PEG/PEGbiotin display distinctive red fluorescence intensity. (c) A magnified view of four ITO microelectrodes: green attributed to fibrinogen/488 and red to streptavidin/633, which was binding specifically to biotinylated PLL-g-PEG/PEGbiotin.

matrix array of 60 ITO electrodes (minus four corner electrodes). Each ITO microelectrode measured 40 × 40 µm2 with an interspacing of 200 µm between them. Figure 1a shows the crosssectional view of the glass-based ITO microarray. Two different layouts for the MEAs were used for the fluorescence microscopy experiments: (1) MEA without the insulating SU-8 (epoxy) coating, which exposed the lead connections this particular MEA used for the adsorption of fluorescently labeled PLL-g-PEG/488 and fluorescently labeled vesicles/488). (2) MEA coated with an insulating SU-8 that covered the lead connection, hence showing only the square microstructure, which is used for the desorption of fluorescently labeled PLL-g-PEG/633 and fluorescently labeled proteins). There is no experimental difference in using both types of substrates for the fluorescence microscopy studies. The MEA biochip was attached onto the printed circuit board by means of screen-printing. The chemicals used to produce the buffer solution were purchased from Fluka AG. The 10 mM 4-(2-hydroxyethyl)-

piperazine-1- ethanesulfonic acid (HEPES) was used as a buffer solution supplemented by 150 mM NaCl. The HEPES solution was adjusted to pH 7.4 using NaOH. All aqueous solutions were prepared using ultrapure water filtered through Milli-Q Gradient A10 filters, purchased from Millipore AG. The TOC of the highpurity water was below 3 ppb. PLL(20)-g[3.6]-PEG(2) was prepared and characterized inhouse according to earlier published protocols.23 The notation refers to a PLL backbone of 20 kDa with PEG side chains of 2 kDa, and g is the grafting ratio (i.e., the number of lysine monomers per PEG side chains). Alexa Fluor-633 carboxylic acid, succinimidyl ester was purchased from Molecular Probes. Labeling of the PLL-g-PEG primarily involved the reaction of the succinimidyl activated acid group on the Alexa Fluor-633 with the free amino groups of PLL using following procedure. Alexa Fluor(23) Kenausis, G. L.; Voros, J.; Elbert, D. L.; Huang, N.; Hofer, R.; Ruiz-Taylor, L.; Textor, M.; Hubbell, J. A.; Spencer, N. D. J. Phys. Chem. B 2000, 104, 3298-3309.

Analytical Chemistry, Vol. 78, No. 3, February 1, 2006

715

Figure 5. Confocal fluorescence images. (a) The MEA that was exposed to fluorescently labeled protein-resistant copolymer PLL-g-PEG/633 for 30 min at ambient conditions and rinsed. (b) An electrical polarization of +1800 mV (referenced to silver electrode) released the polymer from the selected microelectrodes (white arrows) thus exposing the underlying ITO surfaces. (c) Fluorescently labeled streptavidin/633 (red) was introduced and it was observed to adsorb specifically onto the two polarized ITO microelectrodes. (d) The MEA was subsequently exposed to the biotinylated fluorescently labeled liposomes DOPC/488 (green). The vesicles were observed to bind specifically onto the two ITO microelectrodes, which were covered with an adlayer of streptavidin/633. The other regions remained protein-resistant with an adlayer of PLLg-PEG/633 (red).

633 (1 mg) in dry dimethyl sulfoxide (0.1 mL) was added to a solution of PLL (20)-g-[3.6]-PEG(2) (5.4 mg in 1 mL of 0.1 M NaHCO3 buffer, pH 8.3) (in the following to be referred as PLLg-PEG/633). The PLL-g-PEG/633 solution was protected from light and stirred for 6 h at room temperature. PLL-g-PEG/633 was further purified by separating the unreacted Alexa-633 molecules with column chromatography. The sample was loaded on a column (40 cm long, diameter 1.5 cm) packed with P30 Biogel and was eluted with buffer solution (10 mM potassium phosphate, 150 mM NaCl, 2 mM sodium azide, pH 7.2). A similar fluorescently labeled protein-resistant polymer with a different fluorescence excitation wavelength of 488 nm was synthesized according to previous publications,24 and it is referred here as PLL-g-PEG/488. It should be noted that the labeling involves only 1-2 labels/PLL molecule, which has more than 100 free amino groups and as such the labeling does not significantly influence the behavior of the molecule. (Results not shown.) Biotinylated PLL-g-PEG/PEGbiotin was prepared with the graft copolymer PLL-g-PEG carrying the terminal biotin group on 50% of the PEG chains, according to previous publications.17 All the protein-resistant polymers were dissolved in HEPES solution with a concentration of 0.5 mg/mL. Fluorescently labeled human fibrinogen Alexa 488 (fibrinogen/ 488) and streptavidin Texas Red 633 (streptavidin/633) were

obtained from Molecular Probes. The proteins were dissolved in HEPES solution with a concentration of 50 µg/mL. The phospholipids used in our study are as follows: dioleoylphosphatidylcholine (DOPC), dioleoylphosphatidylethanolamine (DOPE), and fluorescently labeled nitrobenzoxadiazole phosphatidylcholine (NBD-DOPC) were purchased from Avanti Polar Lipids Inc. The DOPC/DOPE/NBD-DOPC were mixed with a ratio of 9.6:3:1 (in the following to be referred as DOPC/488) in chloroform and dried with nitrogen to obtain a unilamellar bilayer. The biotinylated fluorescently labeled vesicles, DOPC/ 488, were then prepared according previous publications.25 The labeled vesicles produced were dissolved in the HEPES solution with a concentration of 0.5 mg/ml and subsequently extruded through a filter membrane to obtain a spherical diameter of 50 nm. Methods. An electrochemical flow cell was made from poly(ether ether ketone), and it had an overall outer diameter of 15 mm and a height of 11 mm. The inner diameter of the flow cell was 9 mm and it had a flow chamber volume of 60 µL. The inlet and outlet flow channels through the flow chamber were fitted with Teflon tubing. A three-electrode configuration was adopted for the standard electrochemical cell setup: reference and counter electrodes were 2-mm-diameter 99.9% pure annealed silver and 0.5-mm-diameter

(24) Csucs, G.; Michel, R.; Lussi, J. W.; Textor, M.; Danuser, G. Biomaterials 2003, 24, 1713-1720.

(25) MacDonald, R. C.; et al. Biochim. Biophys. Acta-Biomembr. 1991, 1061, 297-303.

716

Analytical Chemistry, Vol. 78, No. 3, February 1, 2006

99.95% pure platinum (Johnson Matthey & Brandenberger AG), respectively. The reference electrode was tight-fitted through the center of the electrochemical flow cell while the counter electrode was placed inside the Teflon tubing of the outlet flow channels. Both the reference and counter electrodes were connected to the potentiostat via an electrical cable wire. The distance between the reference/counter and working electrode was maintained at 1 mm. The working electrode was the MEA, which was rinsed with ethanol and ultrapure water filtered through the Milli-Q Gradient A10 filters. The biochip was then blown dry with filtered nitrogen and subsequently cleaned within an oxygen-plasma cleaner (PDC32G, Harrick) for 2 min. The electrochemical flow cell was fitted over the MEA by tightsealing the assembly with a Karlrez O-ring and secured carefully onto an electronic multiplexing control unit. The electronic multiplexing control unit enables multiple addressing on the ITO microelectrodes and it was interfaced between the potentiostat and the biochip. The entire setup was then placed over the confocal laser fluorescence microscope for imaging analysis. (1) Selective Surface Desorption of Fluorescently Labeled PLL-g-PEG/633. Labeled protein-resistant polymer PLL-g-PEG/ 633 was first uniformly adsorbed (concentration, buffer) onto the MEA to create a passivating background. The incubation condition for the polymer adsorption onto the MEA surface was 30 min at room temperature (Figure 1b). The flow cell was then rinsed with HEPES solution to remove any excess polymer. Specific ITO microelectrodes were electrically activated at +1800 mV (reference to silver electrode) for 45 min (Figure 1c). This would selectively desorb the protein-resistant adlayer from the electrically addressed ITO microelectrodes, hence regenerating the ITO surface for subsequent molecular immobilization (Figure 1d). HEPES solution was injected again to ensure that no excess polymer remains within the electrochemical flow cell. (2) Selective Surface Adsorption of Fluorescently Labeled PLL-g-PEG/488 (Figure 1i). The procedures for surface passivation and selective electrochemical desorption of PLL-gPEG/633 were conducted similarly to that about. Another different fluorescently labeled protein-resistant polymer, PLL-g-PEG/488, was injected into the flow cell and incubated for 30 min at room temperature. Subsequently, the flow cell was rinsed with HEPES solution to remove excess polymer. The ITO electrodes, which were electronically addressed earlier, were electronically reactivated again at +1800 mV. This was followed closely by electronically addressing three inactivated microelectrodes which were adsorbed with PLL-g-PEG/633 at +1800 mV. (3) Selective Surface Immobilization of Fluorescently Labeled Vesicles/488 (Figure 1ii). For the lipid vesicles, the protocols for surface passivation and selective electrochemical desorption of PLL-g-PEG/633 were conducted similarly as discussed in section 1. The flow cell was exposed to a solution of fluorescently labeled streptavidin/488 for 10 min at room conditions. This was followed by rinsing with HEPES solution. Then the flow cell was exposed a solution of biotinylated fluorescently

labeled lipid vesicles DOPC/488 for 10 min. Finally, HEPES solution was injected to rinse the flow cell. (4) Selective Surface Immobilization of Fluorescently Labeled proteins (Figure 1iii). Biotinylated PLL-g-PEG was uniformly adsorbed onto the MEA and incubated for 30 min at room temperature (Figure 1c). Then the flow cell was rinsed with HEPES solution to remove any excess polymer. Selected ITO microelectrodes were electronically addressed at +1800 mV (referenced to silver electrode) to desorb the functionalized polymer from the ITO surfaces. The flow cell was rinsed with HEPES solution again. This was followed by injecting a solution of fluorescently labeled fibrinogen/488 and incubating it for 10 min. Subsequently, excess proteins were rinsed away with the HEPES solution. Fluorescently labeled streptavidin/633 was introduced into the flow cell and incubated for 10 min. The fluorescently labeled streptavidin/633 is expected to adsorb onto the background and onto the inactivated ITO spots. Finally, the flow cell was rinsed to remove any excess proteins. CONCLUSION This novel work has clearly demonstrated the synergy in using an electrical stimulus to pattern biological moieties onto a microarray. We adopted an electronic multiplexing capability to dynamically address an indium tin oxide microelectrode array that has been uniformly coated with a protein-resistant adlayer, PLLg-PEG/633. A well-defined platform was then made available by electronically addressing individual microelectrodes and hence releasing the polymer from the selected ITO surfaces. By integrating simple surface chemistry onto the surfaces of the bare ITO microelectrodes, lipid vesicles could be arrayed, a major step toward the creation of membrane protein microarrays. An electroactive control within a proteomic biosensor has several benefits as it offers rapid screening and discrimination of different biomolecules with high selectivity. We showed that the most obvious advantage gained from this study is a regenerative platform for a variety of biomolecular immobilization strategies with an electrically switchable surface that can be modified repeatedly to release an adlayer of protein-resistant polymer. Future applications with an addressable electronic microarray are not solely limited within the biosensors field but also foreseeable to include microfluidics, drug delivery, and manipulation of cellular neuron networks for tissue engineering. ACKNOWLEDGMENT We thank Dr. Gabor Csucs, Light Microscopy Center, ETH Zurich, for assisting with the fluorescence microscopy and Fernanda Rossetti, Laboratory for Surface Science and Technology, ETH Zurich, for the valuable discussion. Kind financial support from the Swiss Federal Laboratories for Materials Testing and Research (EMPA) is also acknowledged. Received for review July 12, 2005. Accepted November 9, 2005. AC051244A

Analytical Chemistry, Vol. 78, No. 3, February 1, 2006

717