Carboxyl-Terminated Dendrimer-Coated Bioactive Interface for Protein

(d) Fleer, G. J. Polymers at Interfaces, 1st ed.; Chapman & Hall: London, 1993. There is no corresponding record for this reference. (e) Malmsten, M.;...
0 downloads 0 Views 324KB Size
Download PDF
5670

Langmuir 2007, 23, 5670-5677

Carboxyl-Terminated Dendrimer-Coated Bioactive Interface for Protein Microarray: High-Sensitivity Detection of Antigen in Complex Biological Samples Parayil Kumaran Ajikumar,*,† Jin Kiat Ng,† Yew Chung Tang,† Jim Yang Lee,†,‡ Gregory Stephanopoulos,†,§ and Heng-Phon Too*,†,| MEBCS Program, Singapore-MIT Alliance, National UniVersity of Singapore, 4 Engineering DriVe 3, Singapore 117546, Chemical and Biomolecular Engineering, National UniVersity of Singapore, Singapore 117546, Department of Chemical Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, and Department of Biochemistry, Yong Loo Lin School of Medicine, 8 Medical DriVe, National UniVersity of Singapore, Singapore 117597 ReceiVed December 22, 2006. In Final Form: February 13, 2007 Protein microarrays are promising tools that can potentially enable high throughput proteomic screening in areas such as disease diagnosis and drug discovery. A critical aspect in the development of protein microarrays is the optimization of the array’s surface chemistry to achieve the high sensitivity required for detection of proteins in cell lysate and other complex biological mixtures. In the present study, a high-density antibody array with minimal nonspecific cellular protein adsorption was prepared using a glass surface coated with a poly(propyleneimine) dendrimer terminated with carboxyl group (PAMAM-COOH). The carboxyl-terminated dendrimer-modified surface has almost similar nonspecific cellular protein adsorption when compared to an inert PEG-modified surface. In addition, the multiple functional sites available for reaction on the dendrimer surface facilitated high-density immobilization of antibodies and efficient capture of bioanalytes. Various molecules were tested for their ability to block or deactivate the reactive carboxyl surface after antibody immobilization to further reduce the nonspecific binding. A short oligoethylene glycol (NH2-d4-PEG-COOH), was found to significantly improve the signal-to-noise ratio of the assay, resulting in higher sensitivity. The properties and functional qualities of the various surfaces were characterized by contact angle and AFM measurements. Nonspecific protein adsorption and protein immobilization as a function of dendrimer generations and sensitivity of antigen capturing from a buffer (1 pM) as well as from the complex cell lysate (10 pM) system were examined. Our detailed experimental studies demonstrated a facile method of preparing surfaces with high protein loading and low nonspecific protein binding for the development of high sensitivity protein microarrays.

Introduction The identification and validation of complex proteomes has necessitated the search for new methods to map cellular functions.1,2 Among the various techniques developed, protein, peptide, and small molecule microarray technology has gained attention as a high throughput screening tool that can potentially revolutionize basic and pharmaceutical research.3,4 In a common protein microarray format, antibodies are often immobilized on a substrate surface and are used to capture antigens or analyte molecules.5 Both covalent and electrostatic immobilization strategies have been employed to attach the antibodies onto * Corresponding authors. E-mail: [email protected] (T.H.P.) or [email protected] (P.K.A.). † Singapore-MIT Alliance, National University of Singapore. ‡ Chemical and Biomolecular Engineering, National University of Singapore. § Massachusetts Institute of Technology. | Yong Loo Lin School of Medicine. (1) (a) Leuking, A.; Cahill, D. J.; Mullner, S. DDT: Targets 2005, 10, 789. (b) Aebersold, R.; Mann, M. Nature 2003, 422, 198. (c) Adam, G. C.; Sorensen, E. J.; Cravatt, B. F. Mol. Cell Proteomics 2002, 1, 781. (d) Steinberg, T. H.; Pretty, K.; Berggren, K. N.; Kemper, C.; Jones, L.; Diwu, Z.; Haugland, R. P.; Pattonet, W. F. Proteomics 2001, 1, 841. (e) Washburn, M. P.; Wolters, D.; Yates, J. R. Nat. Biotechnol. 2001, 19, 242. (f) Gygi, S. P.; Rist, B.; Gerber, S. A.; Turecek, F.; Gelb, M. H.; Aebersold, R. Nat. Biotechnol. 1999, 17, 994. (2) (a) MacBeath, G. Nat. Genet. 2002, 32, 526. (b) Chen, G. Y. J.; Uttamchandani, M.; Lue, Y. P. R.; Lesaicherre, M. L.; Yao, S. Q. Curr. Top. Med. Chem. 2003, 3, 705. (c) Lee, Y. S.; Mrksich, M. Trends Biotechnol. 2002, 20, 14. (d) Wilson, D. S.; Nock, F. Angew. Chem., Int. Ed. 2003, 42, 494. (3) Kononen, J.; Bubendorf, L.; Kallioniemi, A.; Barlund, M.; Schraml, P.; Leighton, S.; Torhost, J.; Mihatsch, M. J.; Sauter, G.; Kallioniemi, O. P. Nat. Med. 1998, 4, 844. (c) Michaud, G. A.; Bangham, R.; Salcius, M.; Predki, P. F. DDT: Targets 2004, 3, 238.

chemically modified surfaces.6-8 Ideally, the antibodies should be immobilized onto a substrate at high density with uniform distribution, retaining their specific antigen-binding activities, and remain highly accessible to the antigens to be captured.9 In the protein/antibody/peptide microarray, the major objective is the study of interaction partners. Thus, the key requirements are high-binding capacity of interaction entity yet with low non(4) (a) Zhu, H.; Snyder, M. Curr. Opin. Chem. Biol. 2003, 7, 55. (b) Templin, M. F.; Stoll, D.; Schrenk, M.; Traub, P. C.; Vohringer, C. F.; Thomas, O. J. Trends Biotechnol. 2002, 20, 160. (c) Brown, P. O.; Botstein, D. Nature 1999, 21, 33. (d) Fang, Y.; Frutos, A. G.; Lahiri, J. J. Am. Chem. Soc. 2002, 124, 2394. (e) Min, D. H.; Mrksich, M. Curr. Opin. Chem. Biol. 2004, 8, 554. (f) Shin, I.; Cho, J. W.; Boo, D. W. Comb. Chem. High Throughput Screen 2004, 7, 565. (g) MacBeath, G.; Koehler, A. N.; Schreiber, S. L. J. Am. Chem. Soc. 1999, 121, 7967. (h). Winssinger, N.; Ficarro, S.; Schultz, P. G.; Harris, J. L. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 11139. (i) Salisbury, S. M.; Maly, D.; Ellman, J. J. Am. Chem. Soc. 2002, 124, 14868. (5) (a) MacBeath, G.; Schreiber, S. L. Science 2000, 289, 1760. (b) Kusnezow, W.; Jacob, A.; Walijev, A.; Dieh, F.; Hoheisel, J. D. Proteomics 2003, 3, 254. (6) (a) Xu, Q.; Lam, K. S. J. Biomed. Biotechnol. 2003, 5, 257. (b) Li, Y.; Reichert, M. W. Langmuir 2003, 19, 1557. (7) (a) Anzai, J.; Kobayashi, Y.; Nakamura, N.; Nishimura, M.; Hoshi, T. Langmuir 1999, 15, 221. (b) Williams, R. A.; Blanch, H. W. Biosens. Bioelectron. 1994, 9, 159. (c) Levit-Binnum, N.; Lindner, A. B.; Zik, O.; Eshhar, Z.; Moses, E. Anal. Chem. 2003, 75, 1436. (8) (a) 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. (b) Girish, A.; Sun, H.; Yeo, D. S. Y.; Chen, G. Y. J.; Chua, T.-K.; Yao, S. Q. Bioorg. Med. Chem. Lett. 2005, 15, 2447. (c) Lue, R. Y. P.; Chen, G. Y. J.; Hu, Y.; Zhu, Q.; Yao, S. Q. J. Am. Chem. Soc. 2004, 126, 1055. (d) Lue, Y. P. R.; Chen, G. Y. J.; Zhu, Q.; Lesaicherre, M. L.; Yao, S. Q. Methods Mol. Biol. 2004, 264, 85. (e) Lesaicherre, M. L.; Lue, R. Y.; Chen, G. Y.; Zhu, Q.; Yao, S. Q. J. Am. Chem. Soc. 2002, 124, 8768. (9) Butler, J. E. Methods 2000, 22, 4.

10.1021/la063717u CCC: $37.00 © 2007 American Chemical Society Published on Web 03/28/2007

Detection of Antigen in Biological Samples

specific protein background and low variability. Attempts have been made to address these challenges by optimizing the surface chemistry of such protein arrays. Surfaces developed for protein arrays fall into three classes: (i) two-dimensional (2D) planar glass slides with a uniform layer of linear functional groups including aldehyde, epoxy, or carboxylic esters;10 (ii) threedimensional (3D) gel or membrane-coated surfaces such as polyacrylamide, agarose, and nitrocellulose;11-14 and (iii) supramolecular surfaces that consist of macromolecular coatings such as PEG, BSA, avidin, and dendrimers over 2D glass surfaces.15 On 2D glass surfaces, proteins or antibodies bind either by electrostatic interactions or through covalent linkages that give strong and homogeneous attachments. However, limitations such as rapid evaporation of the liquid environment and close surface contact of the immobilized protein may lead to the denaturation of native structure of the protein. Within the 3D gel or membrane surfaces, proteins bind through physical adsorption and are more likely to retain their native protein conformation. However, the lack of strong and specific binding of proteins to the surface often results in protein leaching and non-homogeneous protein immobilization. In principle, the third group of supramolecular surface coatings has the advantages of both 2D and 3D surfaces mentioned above. Supramolecular structures on a planar surface are generated by coating glass slides with large macromolecules, resulting in an interface that is more compatible for minimizing protein denaturation and promoting specific protein binding on the array. Dendrimers are a unique class of polymers that do not form entangled chains associated with linear polymers and yet possess numerous chain ends that can be easily functionalized.16 Previously, poly(amidoamine) (PAMAM) dendrimers with highly uniform and compact amine linkers have been exploited for various applications in high-density DNA microarrays and protein immobilization studies.17,18 However, there has been no systematic evaluation of the surface chemistry of dendrimers and of the factors which determine nonspecific adsorption of proteins onto dendrimer-coated surfaces. This lack of optimization of dendritic surfaces for minimal nonspecific protein binding has limited their use in protein array applications. In addition, the specific antigen-binding activity of antibodies immobilized on a dendrimer surface in a complex biological solution environment has yet to be reported. Here, we report the fabrication and functional evaluation of a nanostructured monolayer of PAMAM dendrimercoated glass surfaces terminated with multiple carboxyl functional groups for the immobilizing antibody.19 Parameters like surface coatings, nonspecific protein adsorption over the modified (10) Sobek, J.; Schlappbach, R. Pharmagenomics 2004, 32. (11) Arenkov, P.; Kukhtin, A.; Gemmell, A.; Voloshchuk, S.; Chupeeva, V.; Mirzabekov, A. Anal. Biochem., 2000, 278, 123. (12) Rubina, A. Y.; Dementieva, E. I.; Stomakhin, A. A.; Darii, E. L.; Pan’kov, S.; Barsky, V. V. E.; Ivanov, S. M.; Konovalova, E. V.; Mirzabekov, A. D. Biotechniques 2003, 34, 1008. (13) Afanassiev, V.; Hanemann, V.; Wo¨lfl, S. Nucleic Acids Res. 2000, 28, E66. (14) Kersten, B.; Possling, A.; Blaesing, F.; Mirgorodskaya, E.; Gobom, J.; Seitz, H. Anal. Biochem. 2004, 331, 303. (15) Angenendt, P.; Glo¨kler, J.; Murphy, D.; Lehrach, H.; Cahill, D. J. Anal. Biochem. 2002, 309, 253. (16) Bosman, A. W.; Janssen, H. M.; Meijer, E. W. Chem. ReV. 1999, 99, 1665. (17) (a) Benters, R.; Niemeyer, C. M.; Drutschmann, D.; Blohm, D.; Wohrle, D. Nucl. Acids Res. 2002, 30, e10-10. (b) Berre, V. L.; Trevisiol, E.; Dagkessamanskaia, A.; Sokol, S.; Caminade, A.-M.; Majoral, J. P.; Meunier, B.; Francois, J. Nucl. Acids Res. 2003, 31, e88-88. (c) Angenendt, P.; Glokler, J.; Sobek, J.; Lehrach, H.; Cahill, D. J. J. Chromatogr. A 2003, 1009, 97. (d) Trevisiol, E.; Berre-Anton, V. L.; Leclaire, J.; Pratviel, G.; Jean-Caminade, A.-M.; Majoral, P.; Francoisa, J. M.; Meunierb, B. New J. Chem. 2003, 27, 1713. (18) (a) Benters, R.; Niemeyer, C. M.; Wo¨hrle. D. Chembiochem 2001, 2, 686. (b) Pathak, S.; Singh, A. K.; McElhanon, J. R.; Dentinger, P. M. Langmuir 2004, 20, 6075.

Langmuir, Vol. 23, No. 10, 2007 5671

surfaces, antibody/protein immobilization, functional group influences after blocking of the reactive surfaces, influence of additive in the analyte solution to prevent the nonspecific protein adsorption, and sensitivity of detection from complex biological fluids were systematically investigated. Experimental Details Materials and Instrumentations. The amine and carboxylterminated PAMAM dendrimers, N-hydroxysuccinimide (NHS), succinic anhydride, Tris, amino ethoxyethanol, glycine, 4-hydroxy3-nitrobenzenesulfonic acid (HNSA), and bovine serum albumin (BSA) were obtained from Sigma-Aldrich. 1-Ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride (EDC) was purchased from Pierce Chemicals. The short-chain oligoethylene glycols, d4-PEGOH, d4-PEG-COOH, d8-PEG-COOH, and d12-PEG-COOH were purchased from Quanta BioDesign Ltd, USA. The amine and epoxy slides were from Genetix, UK. Cy5 fluorescent mono-reactive dye was purchased from Amersham Biosciences. Mouse monoclonal anti-GFP from Molecular Probes, TRITC-labeled swine anti-rabbit IgG, and rabbit anti-mouse IgG from DakoCytomation (USA) were purified using protein A/G column (Pierce Chemicals) according to manufacturer’s instructions. After IgG purification, buffer exchange was performed using Microcon YM-30 spin filters from Millipore. GFP protein was obtained from Upstate, USA. The hydrophobicity of the slides was measured using contact angle measurement on a Rame Hart model 100. Water (0.2 mL) was dropped onto the substrate, and contact angle was measured using the instrument microscope with a built-in protractor. Laser scanning was carried out using a GenePix 4000B (Axon Instruments CA). The images and intensity of profiles of the spots were measured using a Carl Zeiss LSM 510 laser scanning microscope. BioOdyssey Calligrapher MiniArrayer equipped with solid pin (Bio-RAD Laboratories Inc., CA) was used for printing the antibodies. Preparation of Carboxyl-Terminated Dendrimer-Coated Slides. Carboxyl-terminated PAMAM dendrimer slides were prepared using amino-silylated glass slides. The terminal carboxylic acid of the PAMAM dendrimer was activated with a solution of EDC/NHS (1:1, 100 mM) in 0.1 M MES buffer (pH 6.3), layered onto the amine slides, and incubated for 2 h. The slides were then washed with double-distilled water followed by absolute ethanol and then air-dried. The dried dendrimer slides were kept in Vaccuo until use. Preparation of Succinamic Acid Slides. The succinamic slides were prepared according to previously reported procedure.20 Briefly, amine slides were incubated in a solution of 180 mM succinic anhydride in DMF for 30 min and thereafter transferred into a boiling water bath for 2 min. The slides were then washed in absolute ethanol and then dried under a stream of nitrogen. Preparation of Carboxyl-Terminated PEG Slides. PEG-COOH slides were prepared using epoxy-derivatized slides incubated with 50 mM d4-PEG-COOH in 0.1 M NaHCO3 (pH 9) buffer. After 2 h, the slides were washed with water, followed by absolute ethanol, and then dried under a stream of nitrogen. Activation of Carboxyl-Terminated Acid Slides. The carboxylterminated slides were further activated using EDC/HNSA or EDC/ NHS active ester (1:1; 100 mM) in 0.1 M MES, 0.5 M NaCl (pH 6.3) buffer before protein immobilization. The activated slides were washed with double-distilled water, followed by absolute ethanol, and then air-dried using centrifugation. The slides were stored in desiccators under nitrogen at room temperature until use. Cell Culture and Lysate Extraction. Neuro-2A cells (CCL131) were obtained from ATCC and grown at 37 °C in a 5% CO2 atmosphere in DMEM medium (Gibco/Life Technologies) supplemented with 10% (v/v) fetal bovine serum (Hyclone), 100 units/mL penicillin, and 100 µg/mL streptomycin. At 90% confluence, the (19) (a) Ajikumar, P. K.; NG, J. K.; Lee, J. M.; Stephanopoulos, G.; Too, H. P. Int. J. Nanosci. (in press). (b) Ajikumar, P. K.; Kiat, N. J.; Lee, J. Y.; Too, H. P. Abstr. Pap. Am. Chem. Soc. 2005, 229, U1106-U1106 25-PMSE Part 2. (20) Uttanchandani, M.; Chen, G. Y. J.; Lesaicherre, M. L.; Yao, S. Q. Methods Mol. Biol. 2004, 264, 191.

5672 Langmuir, Vol. 23, No. 10, 2007

Ajikumar et al.

Figure 1. Schematic representation of the direct preparation of PAMAM-carboxyl dendrimer slides. The repeating unit for each generation of the PAMAM is shown using a bracket in the left side of the PAMAM-COOH structure. cells were washed twice with 1× phosphate-buffered saline (PBS) and lysed in a solution of 0.5% Triton X-100 in 1× PBS followed by centrifugation at 13.2 rpm for 10 min. The supernatant containing the cellular proteins was recovered for labeling with Cy5 dyes. Labeling of Cell Lysate and Analyte. Cell lysate, rabbit antimouse IgG, and GFP were labeled with Cy5 mono-reactive dye according to the manufacturer’s protocol. The protein samples were diluted to 1 mg/mL in labeling buffer (0.1 M NaCO3, 50 mM NaCl, pH 9) before incubation with Cy5-NHS dye with intermittent shaking at room temperature for 30 min. The reaction was then quenched for 10 min using Tris-HCl (0.1 M, pH 8). Free dyes were removed using protein desalting columns (Pierce) pre-equilibrated in 1× PBS (pH 7.4). Protein concentration of the purified Cy5-labeled cell lysate was quantified using microBCA assay (Pierce). Immobilization of Antibodies onto Slides by Spotting. Antibodies diluted in printing buffer (10% glycerol, 0.1 M NaHCO3 at pH 8.5) were microspotted (∼600 pL for each spot) onto the activated slides using a BioOdyssey Calligrapher MiniArrayer equipped with solid pin from Bio-RAD Laboratories (USA). The coupling reaction was allowed to proceed overnight at 25 °C in a humid chamber. The arrayed slides were then washed with 0.1% TX-100 in 1× PBS (washing buffer) for 20 min and then incubated in blocking buffer for 4 h. Antigen Capture and Detection. Slides were incubated in a humid chamber at 25 °C for 1 h in the dark with antigen molecules in incubation buffer (0.1% TX-100 in 1× PBS) or cell lysate solution diluted in incubation buffer. The slides were then washed 3 × 10 min with wash buffer and then spun dried. The slides were then scanned, and the results were analyzed using GenePix 3.0 software (Axon Instruments).

Results and Discussion Negatively Charged Surfaces Reduce Nonspecific Protein Binding. In a protein microarray format, the cumulative effect of thousands of proteins in complex biological samples, with varying physicochemical properties, contributes to the background noise and limit the accurate detection of the desired targets. Thus, it is critical to optimize the surface chemistry to enable the efficient coupling of the antibody and preserves the specific antigen-binding activities, while minimizing nonspecific adsorption of other biomolecules.21 Protein adsorption to a surface is known to be driven by the net influence of various interdependent interactions such as van der Waals forces, dipolar or hydrogen bonds, electrostatic forces, and hydrophobic effects.22 Studies on polyelectrolyte multilayers has demonstrated that negatively charged carboxyl-terminated surface chemistry shows low protein adsorption and is independent of the bulk polymeric film properties.23 It was demonstrated that the surface properties/ surface charge play a crucial role in controlling the protein (21) (a) Metzger, S. W.; Lochhead, M. J.; Grainger, D. W. IVD Tech. 2002, 8, 39. (b) Kusnezow, W.; Hoheisel, J. D. J. Mol. Recognit. 2003, 16, 165. (c) Cha, T.; Guo, A.; Jun, Y.; Pei, D.; Zhu, X.-Y. Proteomics 2004, 4, 1965. (22) Haynes, C. A.; Norde, W. Colloids Surf., B 1994, 2, 517. (23) Salloum, D. S.; Schlenoff, J. B. Biomacromolecules 2004, 5, 1089.

Figure 2. Contact angles measured for different generation of PAMAM-COOH coated glass slides. Standard deviation was calculated from multiple readings taken from different areas of the slides.

adsorption properties. In addition, the carboxyl-terminated poly(acrylic acid) films are effective at suppressing adsorption of negatively, neutral, and positively charged proteins as well as the multitude of proteins present in fetal bovine serum (FBS).24 Consistent with this observation, the present study show that carboxyl-terminated PAMAM dendrimers (PAMAM-COOH) display low nonspecific binding of proteins and form highly compact hydrophilic nanostructured surfaces with reactive functional groups present at high density on the surface for covalent coupling of antibodies. In the first set of experiments, we performed a systematic study of the effects of various generations of carboxyl-terminated dendrimer on nonspecific protein adsorption and density of antibodies immobilized. Various generations of carboxyl-PAMAM dendrimer (1.5 to 4.5 generations) were prepared by one-step EDC/NHS coupling to aminemodified glass slides (Figure 1). For comparison, a linear linker succinamic acid (SA) slides was prepared by coupling succinic anhydride to the amine slides generating a surface with similar functional -COOH groups to demonstrate the advantage of these macromolecule coated surfaces as compared to 2D surfaces. The surface homogeneity and the properties of the functionalized surfaces were examined by contact angle measurements. Contact angle studies showed that the different PAMAM-COOH coated surfaces have higher contact angle (Figure 2) than the amine-modified slides (45° ( 3°), indicative of the increase in surface hydrophobicity as a result of the poly(propylenimine) backbone. The contact angle of PAMAM-COOH coated surface decreases with higher generation of dendrimer (G4.5 < G3.5 < G2.5 < G1.5). The hydrophilicity of the surfaces in the order of G4.5 > G3.5 > G2.5 > G1.5, is expected due to a greater density of -COOH functional group present on the surface of the slide coupled with higher generation of PAMAM-COOH. The highest generation of PAMAM-COOH (Gen 4.5) shows similar contact angles to the linear linker succinamic acid slides (48° ( (24) Dai, J.; Baker, G. L.; Bruening, M. L. Anal. Chem. 2006, 78, 135.

Detection of Antigen in Biological Samples

Langmuir, Vol. 23, No. 10, 2007 5673

Figure 3. Tapping mode AFM three-dimensional height images of the (A) succinamic acid and (B) PAMAM-COOH G4.5 slides. In order to demonstrate the difference in the surface topography, the image was plotted with same data scale.

2°). The surface homogeneity and supramolecular structure, of the PAMAM-COOH (Gen 4.5) was examined by determining the surface topography using AFM measurements and compared to the 2D slide surfaces such as succinamic acid (Figure 3). It is anticipated that PAMAM-COOH supramolecular structures are more favorable for preserving the native protein conformation as compared to 2D slide surfaces, which are prone to denaturation of immobilized proteins due to close contact between the proteins and the surface.25 The AFM section analysis showed that the island distribution for PAMAM-COOH was 126 nm and for SA slides was 70 nm with a film thickness of 34 ( 3 nm and 20 ( 2 nm, respectively. In addition, analysis of the surface roughness for PAMAM-COOH (3.8 nm) and SA (5.3 nm) slides indicated that PAMAM macromolecule-coated surface is homogeneous as compared to that for the small molecule succinamic acid. Thus, PAMAM-COOH forms a homogeneous, nanostructured, macromolecular layer on the glass surface, creating a quasithree-dimensional surface.26 To evaluate the contribution of nonspecific protein adsorption onto different modified surfaces in complicated matrixes, we examined the nonspecific binding of Cy5-labeled cell lysate solutions on succinamic acid and different generations of PAMAM-COOH-coated slides. Fluorescence intensity measurements from the adsorbed cell lysate proteins showed that the nonspecific protein adsorption on these surfaces was saturated around 50 µg/mL of cell lysate. The nonspecific protein adsorption on PAMAM-COOH slides was lower than succinamic acid functionalized slides (Figure 4A). When compared with lower generation PAMAM-COOH (G1.5, G2.5, and G3.5) surfaces, the higher generation PAMAM-COOH (G4.5) slides showed less protein adsorption. This indicates that more carboxyl groups on the surface may have a greater effect in reducing nonspecific lysate adsorption.23 In addition, the adsorption of different concentrations of Cy5-labeled cell lysates onto anionic (carboxylate, silica) and cationic (aminated) surfaces revealed that PAMAM-COOH (G 4.5) surfaces have at least 100-fold less nonspecific protein adsorption than the amine dendrimer slides (Figure 4B). Higher background was observed with amine (25) Angenendt, P. DDT 2005, 10, 503. (26) Degenhart, G. H.; Dordi, B.; Schonherr, H.; Vancso, G. J. Langmuir 2004, 20, 6216.

Figure 4. Nonspecific adsorption of Cy5-labeled cell lysate to different surfaces. (A) Succinamic acid and different generation of PAMAM-COOH modified slides. (B) Different chemically modified surfaces compared with PAMAM-COOH.

dendrimer-modified surfaces than on low-density aminated surfaces. A similar observation was reported with polymer multilayers terminating with positively charged functional groups where the amount of protein adsorption was substantially higher as compared to negatively charged surfaces.23 This demonstrates that PAMAM-COOH surfaces are superior to PAMAM-NH2 surfaces in reducing noise resulting from nonspecific protein adsorption. Slides prepared from different batches were tested to verify the reproducibility of this system. To further reduce nonspecific adsorption, it is necessary to minimize nonelectrostatic interactions by the use of poly(ethylene oxide) (PEO)

5674 Langmuir, Vol. 23, No. 10, 2007

Ajikumar et al.

Figure 5. Comparison of the spot homogeneity and corresponding intensity profile of antibody loaded on various surfaces: different generation PAMAM-COOH surfaces and succinamic acid (SA) modified surface. The generation number of PAMAM-COOH is indicated in the figure. All the slides were spotted with 600 pL solution of TRITC-labeled IgG antibody (500 µg/mL) and imaged using laser scanning microscope with same exposure time. As indicated by the three-dimensional representation of the signal distribution, the dendrimer slide reveals sharp and highly homogeneous spots.

or poly(ethylene glycol) (PEG).27-29 It is thought that the hydrophilicity, steric repulsion, and excluded volume effects of PEO segments may contribute to lower protein adsorption.30 Compared with PEG-modified surfaces, PAMAM-COOH slides showed 2-fold higher protein adsorption (Figure 4B). Although the PAMAM-COOH does not eliminate nonspecific protein adsorption completely, our studies demonstrated that a higher generation PAMAM-COOH coated slides were efficiently suppressing the nonspecific protein adsorption. Spot Homogeneity and Antibody Immobilization Efficiency. High surface binding capacity of the activated substrates is necessary for high-density immobilization of antibodies. Higher density of antibodies can increase the signal and improve the detection limit of the array. To covalently immobilize the antibodies to the array, we pre-activated the carboxyl surface with HNSA/EDC. Subsequently, varying concentrations of antibody solutions (600 pL) were deposited by contact printing and incubated overnight at room temperature. Homogeneity and the morphology of the spots were then determined by a laser scanning microscope (LSM), and fluorescence intensity was determined with a fluorescence scanner. As compared to the commonly used NHS activation chemistry, HNSA activation showed a greater rate of aminolysis than hydrolysis, consistent with previous report31 (see Supporting Information, Figure S1). (27) (a) Frederixa, F.; Bonroya, K.; Reekmansa, G.; Laureyna, W.; Campitellia, A.; Abramovb, M. A.; Dehaenb, W.; Maesb, G. J. Biochem. Biophys. Methods 2004, 58, 67. (b) Kingshott, P.; Griesser, H. J. Curr. Opin. Solid State Mater. Sci. 1999, 4, 403. (c) Wang, R. L. C.; Kreuzer, H. J.; Grunze, M. J. Phys. Chem. B 1997, 101, 9767. (d) Chapman, R. G.; Ostuni, E.; Yan, L.; Whitesides, G. M. Langmuir 2000, 16, 6927. (28) (a) Shalaby, S. W. Polymers as Biomaterials; Plenum: New York, 1984. (b) Gombotz, W. R.; Guanghui, W.; Hoffman, A. S. J. Appl. Polym. Sci. 1989, 37, 91. (c) Leckband, D.; Sheth, S.; Halperin, A. J. Biomater. Sci., Polym. Ed. 1999, 10, 1125. (29) (a) Jeon, S. I.; Andrade, J. D. J. Colloid Interface Sci. 1991, 142, 159. (b) Jeon, S. I.; Lee, J. H.; Andrade, J. D.; Degennes, P. G. J. Colloid Interface Sci. 1991, 142, 149. (c) Israelachvili, J. N. Intermolecular and Surface Forces, 2nd ed.; Academic Press: London, 1991. (d) Fleer, G. J. Polymers at Interfaces, 1st ed.; Chapman & Hall: London, 1993. (e) Malmsten, M.; Van Alstine, J. M. J. Colloid. Interface Sci. 1996, 177, 502. (30) Frederixa, F.; Bonroya, K.; Reekmansa, G.; Laureyna, W.; Campitellia, A.; Abramovb, M. A.; Dehaenb, W.; Maes, G. J. Biochem. Biophys. Methods 2004, 58, 67. (31) Aldwin, L.; Nitecki, D. E. Anal. Biochem. 1987, 164, 494.

As shown in Figure 5, a better spot morphology and homogeneity was observed with PAMAM-COOH surfaces as compared to linear linker succinamic acid slides as indicated from the 3D representation and profile of the signal intensities. The higher generation PAMAM-COOH surfaces yielded sharper spots, and the signal distribution is more homogeneous. This result is particularly important since the higher binding density as well as homogeneity within a spot is crucial for reliable quantitative analysis of the captured analyte molecules. Many other studies on protein/peptide microarray also showed that one of the major reasons for poor reproducibility is nonuniformity of spot profiles.5a,17c,32 Recent study on the ring formation in protein microarrays showed that the ring structure results from the transport of protein molecules accumulated at the air/water interface to the perimeter of the droplet on a solid surface.33 This effect becomes more significant as the droplet size decreases. The ring structure can be eliminated by adding competitive surfactants to the protein solution or by designing facile surface reactions for protein immobilization. The present study shows that, using a high-density dendrimer surface and 10% glycerol in 0.1 M NaHCO3 (pH 8.5) as printing buffer, the ring structure can be easily overcome. Fluorescence intensities from different concentration of antibodies arrayed on the slides showed that higher generation PAMAM-COOH (G4.5) surfaces have higher antibody loading as compared to lower generation PAMAM-COOH surfaces (Figure 6). This is consistent with the previous studies where BSA was immobilized on PAMAM-NH2 surfaces.18b PAMAMCOOH (G1.5) coated surfaces exhibited the lowest binding capacity and signal intensity were saturated at an antibody concentration of 500 µg/mL, possibly due to the limited number of linkers available for the antibody binding. In contrast, higher generations PAMAM-COOH surfaces were not saturated, which is likely due to the greater number of binding sites provided by the higher generation quasi three-dimensional PAMAM-COOH surfaces. Because of the superior performance in the nonspecific (32) (a) Houseman, B. T.; Huh, J. H.; Kron, S. J.; Mrksich, M. Nat. Biotechnol. 2002, 20, 270. (b) Haab, B. B. Proteomics 2003, 3, 2116. (33) Deng, Y.; Zhu, X. Y.; Kienlen, T.; Guo, A. J. Am. Chem. Soc. 2006, 128, 2768.

Detection of Antigen in Biological Samples

Langmuir, Vol. 23, No. 10, 2007 5675

Figure 6. (A) Comparison of the antibody binding capacity as a function of increasing concentrations of immobilized TRITC-labeled IgG antibodies (10, 25, 50 100, 250, and 500 µg/mL) on various PAMAM-COOH functionalized surfaces. (B) The fluorescence image of the antibody binding on to various PAMAM-COOH functionalized surfaces The arrow represents the increasing concentration of the antibody spotted.

Figure 7. Nonspecific protein adsorption from Cy5-labeled cell lysate (100 µg/mL) on different surfaces after passivation of the reactive PAMAM-COOH surface.

interaction, surface homogeneity, protein immobilization, a PAMAM-COOH with Gen 4.5 was selected for further studies. Effect of Blocking Agents and Additives in Reducing Nonspecific Protein Adsorption. Compared to other immunoassays such as ELISA, the area of reactive, unoccupied surface on a protein microarray is large, and the deactivation of the activated surface to decrease the nonspecific background signal is a key step. Similar to the protein array substrate development, an optimum surface chemistry can be tailored with a blocking agent to minimize protein adsorption by proper selection of surface charge. A number of inert passivating reagents (e.g., Tris, ethanolamine, oligoethylene glycols, and nonionic detergents) have been included into protein arrays to further reduce nonspecific interactions.5b,9,34 We have examined the effect of a number of reagents when used as passivating reagents for blocking the remaining HNSA ester activities on PAMAM-COOH slides after immobilization of the antibodies (Figure 7). Previous studies have shown that PEG polymer with higher surface densities and long chain length exhibit optimal resistance to nonspecific protein adsorption.29,27,35 This observation has been extended to the use of oligo(ethylene glycol) (OEG) selfassembled monolayers (SAMs) to reduce protein adsorption.36 Contrary to our expectation, the surface modified with a short(34) (a) Szleifer, I. Biophys. J. 1997, 72, 595. (b) McPherson, T.; Kidane, A.; Szleifer, I.; Park, K. Langmuir 1998, 14, 176. (35) (a) Li, L.; Chen, S.; Zheng, J.; Ratner, B. D.; Jiang, S. J. Phys. Chem. B 2005, 109, 2934.

chain PEG with -COOH terminal groups (d4-PEG-COOH, NH2CH2-CH2-(-O-CH2-CH2-)3-O-CH2-CH2-COOH) has the lowest protein adsorption as compared to the longer PEG molecules (d8-PEG-COOH, NH2-CH2-CH2-(-O-CH2-CH2-)7-O-CH2-CH2COOH), and d12-PEG-COOH, NH2-CH2-CH2-(-O-CH2CH2-)11-O-CH2-CH2-COOH). Even though the surface chemistry is similar, the long PEGs (d8-PEG-COOH and d12-PEG-COOH) possibly suffer from the lower immobilization efficiency to the activated PAMAM-COOH surfaces as compared with the small analogue d4-PEG-COOH. Carboxyl-terminated PEG surfaces (d4-PEG-COOH) consistently showed 3-fold less protein adsorption as compared to the hydroxyl terminal PEG with same backbone length (d4-PEG-OH). Note that single carbon chain glycine-terminated with carboxyl group has almost similar protein adsorption as compared to the d4-PEG-OH and 2-fold lower protein adsorption than the aminoethoxy ethanol (Figure 7). In summary, passivating the reactive PAMAM COOH surface with d4-PEG-OH, Tris, aminoethoxy ethanol, glycine, or washing the surface with 1× PBS 0.1% TX did not produce significant reduction in background noise. However blocking with BSA solution was effective in minimizing the protein adsorption. On all surfaces analyzed, blocking with BSA and d4-PEG-COOH significantly reduced the nonspecific protein adsorption. Previous studies have shown that the use of additives in medium can severely affect the specific signal intensities when capturing analyte molecules.5b Classical protein blocking agents such as BSA and milk have been reported to have strong inhibitory effects on signal intensities of captured antigens.37 Thus, the signalto-noise ratio of the antigen capturing on PAMAM-COOH (Gen 4.5) slides blocked with Tris, d4-PEG-OH, and BSA were evaluated. Antibodies (500 µg/mL) were arrayed on the amine reactive surface and incubated overnight. The activated surfaces were then passivated with Tris, d4-PEG-COOH, or BSA. Cy5labeled IgG was used as an antigen to test the capturing efficiency on these slides. The net signal intensities of antigen captured on slides passivated with Tris and BSA were lower as compared to the d4-PEG-COOH (Figure 8A) treated surfaces. In addition, the d4-PEG-COOH passivated surfaces have superior signal(36) (a) Ostuni, E.; Chapman, R. G.; Holmlin, R. E.; Takayama, S.; Whitesides, G. M. Langmuir 2001, 17, 5605. (b) Feldman, K.; Hahner, G.; Spencer, N. D.; Harder, P.; Grunze, M. J. Am. Chem. Soc. 1999, 121, 10134. (c) Pertsin, A. J.; Grunze, M. Langmuir 2000, 16, 8829. (d) Pertsin, A. J.; Hayashi, T.; Grunze, M. J. Phys. Chem. B 2002, 106, 12274. (e) Luk, Y. Y.; Kato, M.; Mrksich, M. Langmuir 2000, 16, 9604. (37) Kaur, R.; Dikshit, K. L.; Raje, M. J. Histochem. Cytochem. 2002, 50, 863.

5676 Langmuir, Vol. 23, No. 10, 2007

Ajikumar et al.

Figure 9. Comparison of the sensitivity of capturing GFP in buffer alone (gray bars) and when spiked in cell lysate (black bars) using anti-GFP antibody (0.5 mg/mL) arrayed dendrimer slides.

Figure 8. (A) Detection (fluorescence intensity graph) of IgG captured with anti IgG arrayed on PAMAM-COOH surfaces passivated with Tris, BSA, and d4-PEG-COOH. (B) Corresponding signal-to-noise graph of 1, 10, 100, and 1000 pM antigen captured from the solution. The signal-to-noise ratio was calculated as the ratio between the mean spot signal intensity and the standard deviation of background intensity. Significant differences in all the data points were calculated using paired two tailed t-test; all the calculated p values were < 0.01.

to-noise ratio (Figure 8B). The detection limit in this particular system was as low as 1 pM (150 pg/mL) with the use of d4PEG-COOH to as the passivating agent. Subsequent studies were conducted with slides passivated with d4-PEG-COOH. Many studies have shown that surfaces can be designed to minimize protein adsorption by proper selection of surface chemistry/surface charge, yet no system has been completely effective at eliminating adsorption altogether as generally observed for all protein “resistant” technologies. In order to further reduce the nonspecific protein adsorption from the cell lysate, we have extended our previous observation that the use of non fat milk as additives, which is superior over the use of many other protein and non protein additives.19 The addition of non-fat milk into the lysate solution further reduces nonspecific adsorption of cellular lysates (see Supporting Information, Figures S2 and S3) onto the PAMAM-COOH modified slides. Higher concentration of non fat milk (10%) as an additive did not interfere with the specific capturing of the analyte but reduced the background (nonspecific binding of proteins) noise by at least 4-fold as compared to 1% milk. Thus, nonspecific protein adsorption from cell lysate can be minimized by a combination of appropriate surface modification, the use of an appropriate blocking agent, and the use of nonfat milk as an additive in the capturing medium. Sensitivity and Capturing Efficiency of PAMAM-COOH (G4.5) Modified Slides. To investigate the practicality of the PAMAM-COOH (G4.5) slides and conditions developed in this study, we tested the effectiveness of the slides in capturing of various concentration of GFP as antigen spiked into 100 µg/mL Cy5-labeled cell lysate solution. Mouse monoclonal anti-GFP antibody (500 µg/mL) was arrayed on PAMAM-COOH slides and incubated overnight, and the slides were blocked with d4PEG-COOH. Varying concentrations of Cy5-labeled GFP antigen (1, 10, 100, 1000, and 10000 pM or 0.000028, 0.00028, 0.0028, 0.028, and 0.28 µg /mL) were then incubated and captured on

the arrays. Compared to the observed spot signals, the background was low, despite the presence of the cell lysate. The measured net signal with and without cell lysate is summarized in Figure 9. The detection limit was 1 pM on PAMAM-COOH G4.5 substrate in the case of antigen captured from the buffer alone which is 1000-fold more sensitive than the succinamic acid and PAMAM-COOH G1.5 modified surfaces (see Supporting Information, Figure S4). Similar detection limit of the antigen capture from the buffer (1 pM) was observed by Wacker et al. for DNA directed immobilization (DDI) antibody microarray format using a more complicated sandwich detection method.38 However, in the presence of cell lysate, the assay was not as sensitive, achieving a detection limit of 10 pM. We are currently exploring strategies to improve the signal by using amplification methods. The tremendous variability in the physicochemical properties of proteins, and consequently, the high-density immobilization of the protein molecules on a substrate that can preserve the native structure, activities, and accessibility to their targets remains a significant challenge in the production of a robust protein microarray. Control of the local molecular environment is one of the critical parameters to obtain an optimum substrate that reduces nonspecific protein adsorption. Here, we have explored quasi-3D structures with the multifunctionality of PAMAMCOOH dendrimer to provide a molecular layer of high-density reactive ligands that are highly efficient in antibody loading. When compared to previously fabricated amine-modified dendrimer surfaces,18b the newly fabricated system has very low nonspecific protein adsorption and low detection limit. The majority of the reported antibody arrays by direct spotting used low ligand density surfaces, thus limiting the amount of antibodies immobilized.6,39 However, observed results showed that sensitivity of the antibody microarray can be improved by the appropriate selection of surface chemistry for increasing density for antibody immobilization, passivation of the reactive surface with short PEG moieties and the inclusion of nonfat milk in the capturing solution. It is likely that the deposition of a biocompatible molecular layer of carboxyl terminated dendrimer molecules with high surface density of immobilized probes minimizes the denaturation, thereby increasing the amount of functional antibody. Consequently, the efficiency of capture and the limit of detection of analytes from the complex biological samples are enhanced. (38) Wacker, R.; Schro¨der, H.; Niemeyer, C. M. Anal. Biochem. 2004, 330, 281. (39) (a) Zhu, H.; Snyder, M. Curr. Opin. Chem. Biol. 2001, 5, 40. (b) Yeo, S. Y. D.; Panicker, R. C.; Tan, L. P.; Yao, S. Q. Comb. Chem. High Throughput Screening 2004, 7, 213.

Detection of Antigen in Biological Samples

Conclusions The present study describes a simple, efficient fabrication of high-density carboxyl dendrimer-based glass substrate for protein microarrays. The performance of the fabricated surface for protein microarrays was evaluated in terms of assay specificity and sensitivity, particularly with complex biological samples. A higher generation PAMAM-COOH-coated surfaces (G4.5) has low nonspecific protein adsorption as compared to its lower analogues or succinamic acid surfaces. A series of passivating agents as well as additives were examined and optimized to further reduce the nonspecific protein adsorption from the cell lysate. In addition, the spot homogeneity of the immobilized antibody improves with higher generations of PAMAM-COOH. The detection limit of antigens was estimated to be about 1 pM for pure analyte

Langmuir, Vol. 23, No. 10, 2007 5677

solution and 10 pM in presence of the cell lysate. In conclusion, our new finding summarizes a single-step preparation of dendrimer modified surfaces with high-density functional sites with low nonspecific protein adsorption and capturing of analytes from cell lysate for antibody microarray application. Acknowledgment. The authors thank the Singapore-MIT Alliance, National University of Singapore for financial support. The first two authors contributed equally to this work. Supporting Information Available: Additional figures as noted in the text. This material is available free of charge via the Internet at http://pubs.acs.org. LA063717U