Peptide Aptamers in Label-Free Protein Detection - American

Protein microarray development is absolutely dependent upon the ability to construct interfaces capable of specific, stable, sensitive, and designable...
0 downloads 0 Views 359KB Size
Download PDF
Anal. Chem. 2007, 79, 1089-1096

Peptide Aptamers in Label-Free Protein Detection: 1. Characterization of the Immobilized Scaffold Jason J. Davis,*,† Jan Tkac,†,‡ Sophie Laurenson,§ and Paul Ko Ferrigno§

Central Research Laboratory, University of Oxford, Mansfield Road, Oxford, OX1 3TA, UK, Institute of Chemistry, Slovak Academy of Sciences, Dubravska cesta 9, 812 37 Bratislava, Slovak Republic, and MRC Cancer Cell Unit, Hutchison/MRC Research Centre, Hills Road, Cambridge, CB2 2XZ, UK

The widespread use of DNA microarray technology for monitoring gene expression has generated valuable insight into various disease states, analyzing clustered gene expression and revealing coregulated gene networks. An analysis of gene expression is not, however, sufficient to provide a knowledge of either protein abundance or function.1 The detection and quantification of protein biomarkers in biological samples lies central to proteomics, drug design, disease prognosis, and therapeutic development.2-4 Though traditional biochemical methods give

invaluable insight into protein presence and function, they are not applicable to high-throughput (potentially massively parallel) protein detection. For this, immobilized microarrays, based on optical detection mechanisms are currently of focus. The generation of viable protein microarrays is, though, challenging. Proteins demonstrate a staggering variety of chemistries, affinities, and specificities, and in many cases, there is requirement for multimerization, formation of partnership with other proteins, or posttranslational modification for activity or specific binding to be significant. There is no simple process for the amplification of proteins (like the polymerase chain reaction in the case of DNA), and expression and purification is neither facile nor does it guarantee functional integrity. The inherent instability of many proteins is a concern in microarray shelf life, and the need to control surface-bound orientation (and thereby maximize the fraction of immobilized protein able to bind target analyte) is demanding.5-8 Though the interaction of protein with a solid surface is, of course, well-studied, this is a highly variable interaction for which a molecular level of understanding is only just emerging.9 In many cases, it appears that initial interactions are predominantly nonspecific, hydrophobic, and reversible. Over time, the protein may unfold and trigger more hydrophobic interactions and, ultimately, irreversible adsorption.10,11 These unfolding interactions are associated with the acquisition of a free energy minimum and, often, a “lateral spreading” of the protein structure.12,13 The designed chemisorption of proteins on surfaces brings with it not only more control of recognition efficacy but may also be associated with increased molecular stability. Simulated and experimental investigations of the effects of surface tethering on both protein fold and these dynamic interfacial interactions has suggested the possibility of large effects, some of which may, in fact, lead to increases in protein stability over the solution-phase form.14 The first protein microarrays were built on the most available and well-characterized biorecognition elementssantibodies. Un-

* To whom correspondence should be addressed. E-mail: jason.davis@ chem.ox.ac.uk. Fax: 0044 1865 275914. Tel: 0044 1865 275914. † University of Oxford. ‡ Institute of Chemistry, Slovak Academy of Sciences. § MRC Cancer Cell Unit. (1) Gygi, S. P.; Rochon, Y.; Franza, B. R.; Aebersold, R. Mol. Cell. Biol. 1999, 19, 1720-1730. (2) Kingsmore, S. F.; Patel, D. D. Curr. Opin. Biotechnol. 2003, 14, 74-81. (3) Mor, G.; Visintin, I.; Lai, Y.; Zhao, H.; Schwartz, P.; Rutherford, T.; Yue, L.; Bray-Ward, P.; Ward, D. C. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 76777682.

(4) Brian, B. H. Proteomics 2003, 3, 2116-2122. (5) LaBaer, J.; Ramachandran, N. Curr. Opin. Chem. Biol. 2005, 9, 14-19. (6) Zhu, H.; Snyder, M. Curr. Opin. Chem. Biol. 2003, 7, 55-63. (7) Sapsford, K. E.; Liron, Z.; Shubin, Y. S.; Ligler, F. S. Anal. Chem. 2001, 73, 5518-5524. (8) Kwon, Y.; Han, Z.; Karatan, E.; Mrksich, M.; Kay, B. K. Anal. Chem. 2004, 76, 5713-5720. (9) Gray, J. J. Curr. Opin. Struct. Biol. 2004, 14, 110-115. (10) Ostuni, E.; Grzybowski, B. A.; Mrksich, M.; Roberts, C. S.; Whitesides, G. M. Langmuir 2003, 19, 1861-1872. (11) Ostuni, E.; Yan, L.; Whitesides, G. M. Colloids Surf., B 1999, 15, 3-30.

Protein microarray development is absolutely dependent upon the ability to construct interfaces capable of specific, stable, sensitive, and designable recognition of specific proteins. Peptide aptamers, being peptide recognition moieties presented and constrained by a robust scaffold protein, offer one possible solution. The relative uniformity of a scaffold protein across potentially many thousands of arrayed peptide aptamers is predicted to simplify the production of microarrays. This paper describes the generation and assaying characteristics of a scaffold protein adlayer. Orientational control of the scaffold protein STM, a triply mutated form of the stable intracellular protein inhibitor stefin A is achieved with a surface cysteine residue, which leads to the presentation of the scaffold recognition surface to solution. Operational stability of the system is excellent, with only a minor decrease in detection sensitivity over time (less than 1% h-1). We use this system to establish a surface plasmon resonance assay offering a limit of detection of 1 nM (150 ng mL-1) and determine the affinity constant of interaction of STM for a cognate antibody to be KD ) 1.47 ( 0.23 nM. Thus, we have established a solid foundation for the future creation of highly multiplexed peptide aptamer microarrays that will be compatible with a broad range of label-free detection technologies.

10.1021/ac061863z CCC: $37.00 Published on Web 01/03/2007

© 2007 American Chemical Society

Analytical Chemistry, Vol. 79, No. 3, February 1, 2007 1089

fortunately, antibodies do not function well in the microarray format, because typically only a small fraction (20%) specifically recognize the target protein with cross-reactivity with other analytes in the microarray format negatively influencing detection selectivity.15,16 Current antibody-based optical assays are commonly based on sandwich assays in which antigen binding to the immobilized antibody is detected through the use of a secondary, labeled, antibody.17,18 Though sensitive, this method is laborious and often requires a specifically labeled secondary antibody for every antigen of interest. The direct electrochemical or optical tagging of analytes also brings significant concern about the degree to which biorecognition is compromised through the use of a “modified” analyte.19 Alternative protein-receptive molecules are thus of considerable interest. DNA/RNA aptamers have been recently used for formation of protein microarrays20-22 and exhibit good chemical/thermal stability and limited batch variation during production. Labeling protocols are flexible and binding properties can be changed on demand. Despite these strengths, only a limited number (∼20) of protein-recognizing DNA/RNA aptamers are currently known23 and those assembled into microarray format have exhibited significant sensitivity toward cations.24 The tuneable detection, high-density surface immobilization and robust structural characteristics of engineered scaffold proteins, called here generically “peptide aptamers”, offer much in the design of microarrays.25 It is, specifically, possible to avoid the considerable problem presented by protein-protein variability by using a single protein to present a large range of interaction surfaces to solution. Engineered scaffold proteins for specific biorecognition events include anticalins modeled on lipocalin structures,26 trinectins derived from a fibronectin III domain,27 and Affibody molecules, which are engineered from the Z domain of protein A.28 The number of peptide aptamers known so far outranks the number of DNA/RNA aptamers. Among peptide aptamers, Affibody technology is the most advanced proteomic tool, and Affibodies have been successfully used for preparation of microarrays using (12) Kondo, A.; Oku, S.; Higashitani, K. J. Colloid Interface Sci. 1991, 143, 214221. (13) Buijs, J.; Hlady, V. J. Colloid Interface Sci. 1997, 190, 171-181. (14) Friedel, M.; Baumketner, A.; Shea, A.; Emma, J. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 8396-8401. (15) Haab, B. B.; Dunham, M. J.; Brown, P. O. Genome Biol. 2001, 2, Research0004. (16) Michaud, G. A.; Salcius, M.; Zhou, F.; Bangham, R.; Bonin, J.; Guo, H.; Snyder, M.; Predki, P. F.; Schweitzer, B. I. Nat. Biotechnol. 2003, 21, 15091512. (17) MacBeath, G.; Schreiber, S. L. Science 2000, 289, 1760-1763. (18) Predki, P. F. Curr. Opin. Chem. Biol. 2004, 8, 8-13. (19) Kirby, R.; Cho, E. J.; Gehrke, B.; Bayer, T.; Park, Y. S.; Neikirk, D. P.; McDevitt, J. T.; Ellington, A. D. Anal. Chem. 2004, 76, 4066-4075. (20) Balamurugan, S.; Obubuafo, A.; Soper, S. A.; McCarley, R. L.; Spivak, D. A. Langmuir 2006, 22, 6446-6453. (21) Radi, A. E.; Sa’nchez, J. L. A.; Baldrich, E.; O’Sullivan, C. K. J. Am. Chem. Soc. 2006, 128, 117-124. (22) LeFloch, F.; Ho, H. A.; Leclerc, M. Anal. Chem. 2006, 78, 4727-4731. (23) Collett, J. R.; Cho, E. J.; Ellington, A. D. Methods 2005, 37, 4-15. (24) Jiang, Y.; Fang, X.; Bai, C. Anal. Chem. 2004, 76, 5230-5235. (25) Ladner, R. C.; Ley, A. C. Curr. Opin. Biotechnol. 2001, 12, 406-410. (26) Skerra, A. Rev. Mol. Biotechnol. 2001, 74, 257-275. (27) Xu, L.; Patti, A.; Gu, K.; Kuimelis, R. G.; Kurz, M.; Lam, T.; Lim, A. C.; Liu, H.; Lohse, P. A.; Sun, L.; Weng, S.; Wagner, R. W.; Lipovsek, D. Chem. Biol. 2002, 9, 933-942. (28) Nord, K.; Gunneriusson, E.; Ringdahl, J.; Stahl, S.; Uhlen, M.; Nygren, P.-A. Nat. Biotechnol. 1997, 15, 772-777.

1090

Analytical Chemistry, Vol. 79, No. 3, February 1, 2007

different immobilization protocols.29 We have recently developed a novel scaffold molecule called STM,30 based on human stefin A, an intracellular inhibitor of cathepsins. Significantly, we have attempted to engineer the scaffold to abolish all interactions with human proteins; this should decrease background binding and, hence, increase the signal-to-noise ratio in any microarray format. STM is also unique in having three sites, distant from each other in the primary sequence of the protein yet adjacent in the folded protein, that are naturally used by stefin A to bind to target proteins. These offer the possibility of creating larger and more complex interaction surfaces with great affinity and specificity. The STM peptide aptamer scaffold (Figure 1) has been successfully used for in vivo studies, but its behavior on surfaces was not known. As a first step toward its use in protein microarrays, we introduce herein the oriented immobilization of our newly developed peptide aptamer scaffold30 on planar gold. Gold surfaces are chemically tuneable and compatible with a range of detection methods, including surface plasmon resonance (SPR), fluorescence, scintillation counting, and electrochemistry.31 Nonspecific interactions can be reduced on backfilling of a receptive surface with an appropriate ethylene glycol, which can be bound to the gold surface by thiol cross-linking.20,32-35 We used STM, which contains no natural thiols, and a newly engineered isoform, called STM-Cys+, with an introduced cysteine residue designed for oriented immobilization on gold. The interfacial properties of STM and STM-Cys+ on gold surfaces were analyzed by SPR and electrochemistry. SPR is a robust technology offering real-time analysis of biorecognition in label-free mode of detection. It is, specifically, sensitive to changes in the index of refraction at or near the surface of a metallic film. Being an in situ technique suitable for the label-free kinetic and thermodynamic analysis of interactions between biomolecules and artificial surfaces, it is readily utilized in optimizing the formation and regeneration of a biorecognition layer.36-38 The key elements of building a stable protein recognition monolayer based on this scaffold, its detection, and regeneration characteristics are addressed here. The technology we have developed will be readily applied to the creation of peptide aptamer (or other protein) microarrays on gold surfaces that will be compatible with a broad range of label-free optical and electrical detection techniques. EXPERIMENTAL SECTION Materials. Glycine and bovine serum albumin (BSA) were purchased from Sigma-Aldrich (Poole, Dorset, UK). An OEG thiol (29) Renberg, B.; Shiroyama, I.; Engfeldt, T.; Nygren, P.-A.; Karlstroem, A. E. Anal. Biochem. 2005, 341, 334-343. (30) Woodman, R.; Yeh, J. T. H.; Laurenson, S.; Ko Ferrigno, P. J. Mol. Biol. 2005, 352, 1118-1133. (31) Houseman, B. T.; Huh, J. H.; Kron, S. J.; Mrksich, M. Nat. Biotechnol. 2002, 20, 270-274. (32) Uchida, K.; Otsuka, H.; Kaneko, M.; Kataoka, K.; Nagasaki, Y. Anal. Chem. 2005, 77, 1075-1080. (33) Seigel, R. R.; Harder, P.; Dahint, R.; Grunze, M.; Josse, F.; Mrksich, M.; Whitesides, G. M. Anal. Chem. 1997, 69, 3321-3328. (34) Klenkar, G.; Valiokas, R.; Lundstroem, I.; Tinazli, A.; Tampe, R.; Piehler, J.; Liedberg, B. Anal. Chem. 2006, 78, 3643-3650. (35) Qian, X.; Metallo, S. J.; Choi, I. S.; Wu, H.; Liang, M. N.; Whitesides, G. M. Anal. Chem. 2003, 74, 1805-1810. (36) Fivash, M.; Towler, E. M.; Fisher, R. J. Curr. Opin. Biotechnol. 1998, 9, 97-101. (37) Malmqvist, M. Nature 1993, 361, 186-187. (38) Cooper, M. A. Nat. Rev. Drug Discovery 2002, 1, 515-528.

Figure 1. Predicted structure of the Cys-STM fusion protein. Panel A. The structure of STM (green) is a homology model produced using Swiss-Model55 derived from the crystal structure of stefin A complexed with cathepsin H.56 STM differs from stefin A by five amino acid residues, and these substitutions are shown in yellow and red. The orange arrow highlights the interaction site created by the insertion of peptide recognition moieties following Pro73, highlighted in red. Shown in blue are the first 51 amino acids of a fusion protein created for optimal purification and immobilization of STM, using the thiol-containing cysteine (Cys) residue highlighted in yellow for chemisorption to gold. No structural information is available for these 51 residues, which are therefore shown as a simple loop. Panel B. Primary sequence and secondary structure prediction for the STM fusion protein. Red bars indicate predicted R-helices, green indicates β-strand, and blue indicates a loop and yellow a coil. The method used correctly predicted the known secondary structure elements of Stefin A shown in (A) and suggests that the region shown as unstructured in (A) may in fact comprise a more robust stalk based on three predicted R-helices (helix 1, residues 1-5; helix 2, residues 16-28’ helix 3, 37-53) separated by turns stretching from residues 6-15 and 29-36.

(1-mercapto-11-undecyl)tri(ethylene glycol), 99+% Lab Grade) was purchased from Asemblon (Seattle, WA). All other chemicals used were of highest purity available without further purification. Running buffer used for SPR measurements consists of 10 mM phosphate pH 7.0. Construction of Prokaryotic Expression Plasmids. The vector pET30a+CYS+ was constructed using a QuickChange Mutagenesis Kit (Stratagene) to introduce a single cysteine residue between the amino-terminal hexahistidine (HIS6-tag) and the multiple cloning site. STM was subcloned into pET30a+ and pET30a+CYS+ vectors in the EcoRI restriction site to create STMpET30a+ and Cys+STM-pET30a+ plasmids to express STM without or with an amino-terminal cysteine residue. Expression of Recombinant STM and STM-CYS+ Protein. STM and STM-Cys+ were expressed as HIS6-tag fusion proteins in BL21(DE3)pLysS Escherichia coli cells. The cells containing STM-pET30a+ and STM-pET30a+Cys+ plasmids were grown to mid-log phase at 37 °C and induced with 1 mM isopropyl β-Dthiogalactopyranoside at 37 °C for 3 h. Cells were harvested by

centrifugation and lysed for 20 min at room temperature with BugBuster primary-amine-free Bacterial Lysis Reagent (Novagen, Nottingham, UK) including protease inhibitors (Roche, Lewes, UK). The lysate was centrifuged at 4000g for 30 min and incubated with nickel-nitrilotriacetic acid (Ni-NTA) resin (Qiagen, Crawley, UK) for 2 h at 4 °C. The Ni-NTA resin was washed with 30 volumes of wash buffer (50 mM sodium phosphate, 300 mM sodium chloride, 20 mM imidazole, pH 7.4). HIS6-tagged fusion protein was eluted from the resin with elution buffer (50 mM sodium phosphate, 300 mM sodium chloride, 150 mM imidazole, pH 7.4). The eluted protein was exchanged into phosphatebuffered saline (PBS) using PD-10 protein desalting columns (GE Healthcare). Protein purification was analyzed by SDS-PAGE and Western blotting with anti-human cystatinA monoclonal antibody (AbCam, Cambridge, UK). Immobilization Protocol. A gold chip (Windsor Scientific UK Ltd.) was either used as received or treated by a hot piranha (1 + 3 mixture of concentrated H2SO4 + concentrated H2O2; handle with special care) for 2 min before modification. STM-Cys+ protein Analytical Chemistry, Vol. 79, No. 3, February 1, 2007

1091

(10 µM in 10 mM phosphate buffer pH 7.0) was mixed with a tris(2-carboxyethyl)phosphine hydrochloride (TCEP) gel (TCEP immobilized on a dextran matrix, Pierce) and allowed to react for 45-60 min at room temperature to selectively reduce any interprotein and disulfide bonds and maximize the availability of free sulfhydryl groups for immobilization. The protein was then centrifuged (1000g for 1 min) and supernatant collected and immediately used. Where STM protein was immobilized on a gold surface via physisorption rather than thiol-mediated chemisorption, the TCEP gel treatment was omitted. In order to obtain a biorecognition surface with low nonspecific binding, OEG thiol (either 0.8 mM or 8 µM in 10 mM potassium phosphate buffer pH 7.0) was backfilled on the STM-Cys+ layer for 20 min and extensively washed by the running buffer. Both immobilization of the STM protein and backfilling was performed in situ, with the gold chip mounted in the spectrometer. The TCEP gel pretreatment protocol for preparation of STMCys+ (formation of free sulfhydryl groups) protein described above was used for immobilization of STM-Cys+ protein on polished polycrystalline gold electrodes (Bioanalytical Systems, West Lafayette, IN, diameter of 1.6 mm), from a 1 µM solution of STMCys+ protein incubated statically for 1 h. The surface of the electrode was carefully washed with distilled water, and electrodes were ready for backfilling by OEG-thiol (8 µM, 1 h). Electrodes were again carefully washed with deionized water before further use. SPR Assays. For all measurements, a double-channel SPR reader was used (Eco Chemie). The instrument is equipped with a cuvette and a gold sensor disk (diameter 17 mm) is mounted to the optical lens through index-matching oil. An autosampler is used to inject or remove test solutions, and measurements were typically carried out in the initial volume of running buffer of 50 µL, with sample injections of 5 µL. Measurements were performed under stirred conditions, where the solution was continuously aspirated and dispensed at a flow rate of 17 µL s-1. The SPR response is expressed as an angle shift of a minimum of surface plasmon resonance in millidegree. Kinetic evaluation software was used to obtain kinetic constants of interaction of immobilized STMCys+ protein with its model target protein. For these experiments, we utilized a monoclonal anti-STM antibody (Abcam, Cambridge, UK). The following solutions were tested for their ability to regenerate the STM-Cys+ surface after target protein binding: 10 mM HCl, 10 mM NaOH, and 10 mM glycine-HCl buffer with pH values ranging from 2.0 to 3.5. The absolute amount of bound protein was determined using the relationship 122 mdeg ) 1 ng mm-2.39 Electrochemistry. Electrochemical characterization (blocking properties, reductive desorption, capacitance detection) was performed using a potentiostat µAutolab Type II (Windsor Scientific UK Ltd.). Where gold electrodes were used, a saturated calomel reference electrode together with a platinum wire counter electrode were used. Macrodisk gold electrodes were pretreated by reductive desorption in 0.1 M NaOH in deionized water (20%) in methanol (80%) by cyclic voltammetry from -0.5 to -1.5 V at a sweep rate of 100 mV s-1 (10 scans). Electrodes were rinsed with deionized water and treated with hot piranha for 20-25 min and then rinsed in water, polished on 1- and 0.1-µm alumina/ (39) Su, X.; Wu, Y.-J.; Knoll, W. Biosens. Bioelectron. 21, 719-726.

1092

Analytical Chemistry, Vol. 79, No. 3, February 1, 2007

diamond slurry (Struers A/S), and sonicated in water for 3 min. Blocking properties of gold/modified gold were analyzed using 10 mM ferricyanide in 1 M KCl. Reductive desorption experiments were run in 100 mM NaOH at sweep rates of 10 mV s-1. Capacitance changes were monitored using chronoamperometry by stepping a working potential from 0 to 0.05 V40 under N2. Atomic Force Microscopy. Ambient tapping mode atomic force microscopy (TMAFM) imaging was carried out with a MultiMode microscope (Digital Instruments, Ltd.) in conjunction with a Nanoscope IIIa control system. Both E and J scanners were used, with lateral ranges of ∼10 and ∼125 µm, respectively. Samples were prepared by incubating STM-Cys+ (1 µM in 10 mM PBS pH 7.0) with a freshly annealed gold on mica surface overnight at 4 °C. The gold surfaces were washed with 10 mM PBS buffer pH 7.0 and 10 mM glycine buffer pH 2.5 prior to imaging. Etched Supersharp silicon probes, attached to triangular cantilevers 100-200 µm in length (Nanosensors), were operated at resonances in the 275-330 kHz range (drive 400-1500 mV). Height, amplitude, and phase data were simultaneously collected, the latter with a Digital Instruments phase extender module. Before imaging, scanner calibration in the xy plane was checked through the use of a 1-µm calibration grid (LOT Oriel, Surrey, U.K.) and in the z direction by imaging atomic steps on a cleaved highly orientated pyrolytic graphite surface. Samples were engaged at zero scan size to minimize possible contamination (biofouling) of the probe. Topographic data were regularly recorded simultaneously in trace and retrace to check for scan artifacts. Scan rates were 1-5 Hz. RESULTS AND DISCUSSION Tethered versus Nonspecific Scaffold Immobilization. Both peptide aptamer scaffold proteins (STM-Cys+ and STM) were immobilized on a clean gold surface in situ by injection of protein solutions to a final concentration of 1 µM. SPR binding curves showed less of the native STM protein was immobilized (452 ( 50 mdeg) compared to the engineered STM-Cys+ protein (770 ( 24 mdeg) (Figure 2A). Repeated (typically five) washings of the surface with the regeneration buffer (see below) reduced these molecular coverages to 284 ( 27 mdeg for STM (equivalent to a protein density on the surface of 2.3 ( 0.2 ng mm-2 or 23 ( 2 pmol cm-2) and 597 ( 16 mdeg (equivalent to 4.9 ( 0.1 ng mm-2 or 49 ( 1 pmol cm-2) for STM-Cys+. These observations are ascribed to the removal of loosely bound protein prior to the establishment of a stable baseline and a monolayer of either physisorbed (STM) or chemisorbed (STM-Cys+) STM protein. Reductive stripping voltammetry of the STM-Cys+ adlayer confirmed the presence of an anchoring thiolate-gold bond and indicated a molecular coverage of 51 ( 5 pmol cm-2, a figure in good agreement with the plasmon resonance data. The reproducibility and homogeneity of the receptive STM-Cys+ adlayer was further confirmed by direct AFM imaging (Figure 3). In comparing the adsorption kinetics of STM and STM-Cys+ on a bare gold surface (Figure 2B), it is clear that the physisorbed adlayer forms more rapidly. This observation, which has been made in other recent work,41 is likely to be related to the interfacial electron transfer that accompanies chemisorption through the (40) Berggren, C.; Johansson, G. Anal. Chem. 1997, 69, 3651-3657. (41) Pyun, J. C.; Kim, S. D.; Chung, J. W. Anal. Biochem. 2005, 347, 227-233.

Figure 2. Immobilization of STM-Cys+ (solid line) and STM (dashed line) with injected 1 µM protein solution in 10 mM phosphate buffer pH 7.0 (point 1) on a bare SPR gold chip. At point 2, 10 mM phosphate buffer pH 7.0 was injected to start a dissociation phase (A). Kinetics of immobilization of STM-Cys+ (solid line) and STM (dashed line) on a bare SPR gold chip (100%, signal at the end of association phase), arrows points 1 and 2 have the same meaning as in (A) (B). Response toward addition of 6 nM TP (point 1) on a STM-Cys+ surface (solid line) and a STM surface (dashed line). At point 2, draining of the cell followed by injection of 10 mM phosphate buffer pH 7.0 was carried out. The biorecognition surface was regenerated by injection of 10 mM glycine-HCl pH 2.5 at point 3, followed by further draining of the cell and injection of 10 mM phosphate buffer pH 7.0 at point 4 (C). Response toward 6 nM TP on the STM surface showing the contribution of a specific TP binding (a) and nonspecific binding (interaction with bare gold (b). The arrows have the same meaning as in (C) (D). In this study, the SPR gold chip surface was not blocked by OEG thiol after immobilization of proteins (STM or STM-Cys+).

solvent exposed thiol in the case of STM-Cys+. Specifically, the oxidative adsorption is associated with a drop in open circuit potential of the underlying electrode surface. This, in turn, disfavors further chemisorption, although this can be overcome through the application of a positive surface potential.42 Our ultimate goal will be to immobilize peptide aptamers in the STM scaffold, and use them to detect analyte target proteins in solution. To model this process, we have used as a model analyte target protein (TP) an anti-cystatin A antibody, which recognizes a sequence preserved in STM. The use of this antibody has the advantage of confirming that the protein immobilized on the surface is indeed the scaffold protein, rather than a (contaminating) thiol-containing protein of bacterial origin. Note that this assay differs from traditional immunoassays where the antibody is immobilized on a surface. Here, the immobilized detector is the peptide aptamer scaffold protein, and the antibody is a soluble (42) Praig, V. G.; Hall, E. A. H. Anal. Chim. Acta 2003, 500, 323-336.

analyte. Accordingly, the TP was then injected to a final concentration of 6 nM over the biorecognition STM/gold surface. The amount of TP subsequently bound by the STM-Cys+ layer (39.2 mdeg) is approximately double that observed on the STM layer (19.5 mdeg) (Figure 2C). Regeneration of the surface after TP binding (see below) is effective only on the STM-Cys+ layer; with the nonspecifically bound receptive STM layer, a residual amount (14.8 mdeg) of TP remains present on the surface after regeneration (Figure 2D). It is likely that these observations reflect the presence of a degree of TP-bare gold association in the case of the receptive STM layer (where scaffold molecular packing is appreciably lower than with the cysteine-modified protein) (Figure 2A) and are consistent with recent studies highlighting the importance of oriented receptor immobilization.43 Even though nonspecific immobilization procedures might maintain functional (43) Peluso, P.; Wilson, D. S.; Do, D.; Tran, H.; Venkatasubbaiah, M.; Quincy, D.; Heidecker, B.; Poindexter, K.; Tolani, N.; Phelan, M.; Witte, K.; Jung, L. S.; Wagner, P.; Nock, S. Anal. Biochem. 2003, 312, 113-124.

Analytical Chemistry, Vol. 79, No. 3, February 1, 2007

1093

Figure 3. Ambient TMAFM topographic map (500 × 500 nm) of the receptive STM-Cys+ scaffold adlayer. Each feature is assigned to a tip-broadened single molecule.

protein integrity and specific recognition,17 the advantages of using fusion tags that facilitate the generation of a more homogeneous biorecognition layer are clear.44 Blocking versus Backfilling with OEG Thiol. For practical applications, it is very important to design a surface at which nonspecific interactions are minimal. In conventional immunoassays, a blocking agent such as BSA or low fat milk is used.45 Blocking of the STM-Cys+ layer by BSA led here to a decrease in sensitivity of TP detection by 36%. This observation may be related to the relative geometric sizes of the scaffold protein (10 kDa) and BSA (67 kDa). In backfilling exposed components of the gold surface with an octanethiol self-assembling monolayer (SAM) after STM-Cys+ immobilization, nonspecific protein interactions were observed to markedly increase. The strong association of protein with SAMs terminating in hydrophobic moieties is known10,11,46 and reported to decrease as hydrophilicity increases.47 In contrast, backfilling of the receptive STM-Cys+ adlayer with OEG thiol (which forms, in isolation, an adlayer with a surface coverage of 245 mdeg, corresponding to 2.0 ng mm-2 or 3.6 × 1014 molecules cm-2 in a good agreement with a calculated expectancy),48 resulted in the production of a stable SPR baseline and a marked decrease (44) Zhu, H.; Bilgin, M.; Bangham, R.; Hall, D.; Casamayor, A.; Bertone, P.; Lan, N.; Jansen, R.; Bidlingmaier, S.; Houfek, T.; Mitchell, T.; Miller, P.; Dean, R. A.; Gerstein, M.; Snyder, M. Science 2001, 293, 2101-2105. (45) Vostiar, I.; Tkac, J.; Mandenius, C.-F. Anal. Biochem. 2005, 342, 152-159. (46) Lahiri, J.; Isaacs, L.; Tien, J.; Whitesides, G. M. Anal. Chem. 1999, 71, 777-790. (47) Sigal, G. B.; Mrksich, M.; Whitesides, G. M. J. Am. Chem. Soc. 1998, 120, 3464-3473. (48) Love, J. C.; Estroff, L. A.; Kriebel, J. K.; Nuzzo, R. G.; Whitesides, G. M. Chem. Rev. 2005, 105, 1103-1169.

1094 Analytical Chemistry, Vol. 79, No. 3, February 1, 2007

Table 1. Sensitivity of the Biorecognition Layer with Different Composition toward Target Protein (TP) and toward a Nonselective Probe, BSAa BSA TP selectivity ratio (mdeg/mM) (mdeg/mM) (TP/BSA) STM-Cys+/C8 thiol STM-Cys+ STM-Cys+/OEG thiol a

5.5 1.6 0.025

9690 6138 6980

1762 3836 279 200

Molar selectivity ratio is shown as well.

in nonspecific interactions. Specifically, the OEG-thiol surface was observed to be virtually insensitive to BSA addition up to concentrations of 10 µM (Table 1) with the selectivity ratio for TP/BSA recognition increasing 73-fold from 3836 at the STMCys+ surface to 279 200 on the STM-Cys+/OEG thiol surface. Importantly, the sensitivity of TP detection was not found to be detectably influenced by this backfilling process. The response toward 10 nM TP is 63.5 mdeg for the STM-Cys+ layer before backfilling with OEG thiol and 62.1 mdeg after OEG thiol backfilling, and approximately the same decrease in sensitivity was observed in a reference channel with only STM-Cys+ immobilized, indicating that the OEG thiol adlayer thickness (∼2.22.4 nm34,35) does not inhibit exposure of the scaffold protein to solution or TP. The formation of a biorecognition layer on gold electrodes was followed using a redox probesferricyanidesafter every building step. Redox probe blocking experiments indicate that diffusive access to the underlying gold is only perturbed to a minor degree

Figure 4. Blocking properties of bare gold electrode surface (solid line), a gold electrode surface modified statically by 1 µM of STMCys+ protein for 1 h at 4 °C (dashed line), and a gold electrode surface modified by STM-Cys+ (1 µM for 1 h) and backfilled statically by 8 µM OEG thiol for 1 h (dotted line). CV was taken in the presence of 10 mM ferricyanide in 1 M KCl solution at a sweep rate of 50 mV s-1.

Figure 5. Stability of the SPR sensor response through repeated injection (6 times) of TP (7 nM). The surface is regenerated with 10 mM glycine-HCl pH 2.5 (injected over the surface for 2 min) prior to each repeat. The dashed line shows the SPR signal after TP injection (signal after TP injection substracted from the baseline signal before TP injection), while the solid line shows absolute changes of the SPR signal before TP injection.

by the scaffold adlayer but is essentially blocked on OEG backfilling (Figure 4). Chronoamperometric experiments revealed a change in the interfacial capacitance of the gold surface with every modification step; specifically, capacitance fell stepwise from 15.6 ( 0.7 µF cm-2 for bare gold, to 11.3 ( 1.5 µF cm-2 with the formation of STM-Cys+ adlayers, to 5.7 ( 0.9 µF cm-2 with OEG thiol backfilling. For comparison, the interfacial capacitance of a pure OEG layer (8 µM solution for 1h) is 4.1 ( 0.6 µF cm-2. These values are in good agreement with previously published results.49,50 Regeneration Conditions. The recognition layer discussed thus far can be regenerated after TP binding by injection of a suitable regeneration buffer. Such solutions, in an ideal case, should be capable of removing the target protein from the surface (49) Lee, K. A.; C., B. S. Langmuir 1990, 6, 709-712. (50) Johnson, P. A.; Levicky, R. Langmuir 2004, 20, 9621-9627.

Figure 6. SPR sensorgrams for injection of TP (2.5, 5, 10, and 40 nM) performed on the STM-Cys+/OEG thiol surface (A). A calibration curve for detection of TP on STM-Cys+/OEG thiol surface (solid line) and on STM-Cys surface (dashed line) (B). In both cases, SPR response was read at the time interval shown by a vertical dashed line in (A). The sensorgrams were obtained by running SPR experiment in 10 mM phosphate pH 7.0 and using 10 mM glycine-HCl buffer pH 2.5 injected for 2 min to regenerate the surface.

completely, while preserving the responsiveness of the biorecognition layer.45,51 On screening a variety of solutions, including 10 mM HCl, 10 mM NaOH, and 10 mM glycine-HCl buffers with pH values ranging from 2.0 to 3.5, the best results were obtained with 10 mM glycine pH 2.5 injected over the TP-STM-Cys+ biorecognition pair for 2 min followed by rinsing with running buffer. Three injections of this regeneration solution are typically needed to achieve a stable baseline and biorecognition signal toward 7 nM TP (RSD of 0.5% for n ) 4, Figure 5). The small decrease in signal (1% h-1) underlines both the high stability of the biorecognition layer and the effectiveness of regeneration. Properties of the STM Peptide Aptamer Scaffold. Immobilization of STM-Cys+ on gold is highly reproducible (597 ( 16 mdeg; RSD 2.7%) on freshly cleaned gold surfaces and noticeably lower on those that were not freshly prepared. When building a STM-Cys+/OEG thiol layer on a new gold chip with freshly prepared protein and thiol solutions, TP sensitivity was 6.44 ( 0.36 mdeg nM-1. The receptive scaffold layer can thus be prepared with a high level of reproducibility (RSD of 5.6%), a feature important for future array application. (51) Vostiar, I.; Tkac, J.; Mandenius, C.-F. Anal. Biochem. 2003, 322, 156-163.

Analytical Chemistry, Vol. 79, No. 3, February 1, 2007

1095

The relatively narrow linear range of TP detection in the case of the STM-Cys+ layer alone (without backfilling) is likely to be attributable to the presence of nonspecific TP-bare gold surface interactions (Figure 6). These interactions are strong,46 not reversible through the use of regeneration buffers, and result in an increasing baseline signal when binding experiments are carried out with TP at levels of >20 nM. A signal difference for TP between those two surfaces (STM-Cys+/OEG thiol vs STMCys+) at higher concentration (20 and 40 nM) corresponds to an increase in baseline before injection of TP on the STM-Cys+ surface rather than to a specific recognition event on the STMCys+ surface. In contrast, evaluation of the interaction of TP with the STM-Cys+/OEG thiol surface (Figure 6) revealed kinetic constants ka ) (6.84 ( 0.36) × 104 M-1 s-1 and kd ) (1.01 ( 0.12) × 10-4 s-1 and affinity constants KA ) (6.79 ( 1.06) × 108 M-1 and KD ) (1.47 ( 0.23) × 10-9 M. This KD is in the range of affinity constants for antibodies (typically 0.5-2.0 nM)52,53 showing that the binding of STM to its target protein (antibody) is not compromised by its surface immobilization. For comparison affinity constants of DNA aptamers for their model target protein, lysozyme, are in the range KD ) 29 - 230 nM and can increase to 1.24 µM upon fluorescent labeling.19 The 1 nM (150 ng mL-1) limit of detection demonstrated here, and capped by the method of detection rather than the biorecognition elements,54 should be improvable, potentially by orders of magnitude, on taking this characterized recognition layer to an optical detection platform. CONCLUSION The demands associated with the immense variability of proteins, each ideally requiring independent characterization once immobilized, can be bypassed through the use of a single “tuneable” scaffold protein. We demonstrate herein that, through (52) Katsamba, P. S.; Navratilova, I.; Calderon-Cacia, M.; Fan, L.; Thornton, K.; Zhu, M.; Vanden, Bos, T.; Forte, C.; Friend, D.; Laird-Offringa, I.; Tavares, G.; Whatley, J.; Shi, E.; Widom, A.; Lindquist, K. C.; Klakamp, S.; Drake, A.; Bohmann, D.; Roell, M.; Rose, L.; Dorocke, J.; Roth, B.; Luginbuehl, B.; Myszka, D. G. Anal. Biochem. 2006, 352, 208-221. (53) Svitel, J.; Balbo, A.; Mariuzza, R. A.; Gonzales, N. R.; Schuck, P. Biophys. J. 2003, 84, 4062-4077. (54) Baker, K. N.; Rendall, M. H.; Patel, A.; Boyd, P.; Hoare, M.; Freedman, R. B.; James, D. C. Trends Biotechnol. 2002, 20, 149-156. (55) Schwede, T.; Kopp, J.; Guex, N.; Peitsch, M. C. Nucleic Acids Res. 2003, 31, 3381-3385. (56) Jenko, S.; Dolenc, I.; Guncar, G.; Dobersek, A.; Podobnik, M.; Turk, D. J. Mol. Biol. 2003, 326, 875-885.

1096

Analytical Chemistry, Vol. 79, No. 3, February 1, 2007

the controlled immobilization of a newly developed peptide aptamer scaffold protein, a highly reproducible oriented proteinreceptive surface with nanomolar limits of detection is generated. Regeneration conditions have been established that enable repeated (at least 10 times) use of these layers with high operational stability. Kinetic analyses show that the affinity between the scaffold protein and a monoclonal antibody model target protein is in the range typical for antibody-antigen interaction, even after immobilization of the scaffold protein on gold. Selectivity for the target protein is good and can be enhanced 70-fold through the use of a backfilling thiolated OEG coadsorbate. Peptide aptamers are traditionally selected by yeast two hybrid screening of relatively low complexity peptide aptamer libraries, comprising typically 106-107 potential binders. This method has the advantage that the target protein (typically of human, i.e., eukaryotic origin) is expressed in a eukaryotic (yeast) cell and is more likely to be folded in a biologically relevant manner. The alternative method of screening by phage display allows selection from libraries of much greater complexity (1012-1015) but involves proteins expressed in a prokaryotic cell, where folding may be compromised. We predict that microarrays based on the work described here will allow in vitro screening of arrayed libraries of high complexity, or iterative screening of low complexity microarray libraries, against the multiple proteins and protein isoforms present in lysates of human cells where the native protein confomations are preserved. We anticipate that the solid foundations laid by this work will aid the creation of highly multiplexed peptide aptamer microarrays interfaceable with a broad range of label-free optical or electronic assays. ACKNOWLEDGMENT Work in the PKF laboratory is supported by a Grant-in Aid from the Medical Research Council. SL acknowledges the award of a Scholarship from the Commonwealth Commission. The authors acknowledge financial support from the BBSRC (BB/ D523094/1) and thank Mr. Daniel Axford for acquiring AFM data.

Received for review October 4, 2006. Accepted November 17, 2006. AC061863Z