The Boundary Molecules in a Lysozyme Pattern Exhibit Preferential

Aug 13, 2008 - Polyclonal anti-lysozyme antibodies can bind to the immobilized lysozyme ... model system to investigate the protein pattern/antibody b...
0 downloads 0 Views 2MB Size
Download PDF
10334

Langmuir 2008, 24, 10334-10339

The Boundary Molecules in a Lysozyme Pattern Exhibit Preferential Antibody Binding Pei Gao and Yuguang Cai* Department of Chemistry, UniVersity of Kentucky, Rose Street, Lexington, Kentucky 40506 ReceiVed April 1, 2008. ReVised Manuscript ReceiVed June 27, 2008 Lysozyme was immobilized on a prefabricated carboxylic acid terminated chemical template, forming a tightly packed, one monolayer thick lysozyme pattern. Polyclonal anti-lysozyme antibodies can bind to the immobilized lysozyme pattern. Atomic force microscope (AFM) observation reveals that the antibodies bind to the lysozyme molecules on the pattern edge before they bind to the lysozyme molecules in the pattern interior. Better spatial accessibility and flexibility of the lysozyme molecules on the pattern edge are used to explain the observed antibody binding preference. The topographies of the lysozyme pattern also affect the antibody binding. The antibodies bind to the edge lysozyme from the top if the lysozyme pattern is half-buried in a 10 Å deep channel, whereas the antibodies bind to the edge lysozyme from the side if the lysozyme pattern is immobilized on a protruding terrace. The observed “edge effect” suggests that, for the same protein coverage, reducing the protein pattern feature to the nanoscale will improve the overall binding activity of the immobilized protein toward the antibody.

I. Introduction Protein sensors, protein motors, and bioethanol production demand efficient, stable, and economical protein immobilization methods.1-15 The activity of an immobilized protein is the most important parameter in evaluating a protein immobilization method. Because protein functions are highly site-specific and directional, the dynamical structure and spatial accessibility of protein molecules affect their activities. The spatial accessibility of an immobilized protein molecule is influenced by factors such as adsorption orientations and adsorption sites. The spatial confinements also suppress the relaxation of surface-immobilized proteins.15 These factors cause the activity of the immobilized protein to be different from its solution state. Therefore, to characterize the activity of an immobilized protein, the adsorption structure of the protein has to be understood at the molecular * To whom correspondence should be addressed. E-maio: ycai3@ uky.edu. (1) MacBeath, G.; Schreiber, S. L. Science 2000, 289, 1760. (2) Lee, K. B.; Park, S. J.; Mirkin, C. A.; Smith, J. C.; Mrksich, M. Science 2002, 295, 1702. (3) Lim, J. H.; Ginger, D. S.; Lee, K. B.; Heo, J.; Nam, J. M.; Mirkin, C. A. Angew. Chem., Int. Ed. 2003, 42, 2309. (4) Willner, I.; Katz, E. Angew. Chem., Int. Ed. 2000, 39, 1180. (5) Liu, G. Y.; Amro, N. A. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 5165. (6) Mandana, V.; Zareie, M. H.; Zhang, M. Q. Langmuir 2002, 18, 6671. (7) Rinaldi, R.; Biasco, A.; Maruccio, G.; Arima, V.; Visconti, P.; Cingolani, R.; Facci, P.; De Rienzo, F.; Di Felice, R.; Molinari, E.; Ph. Verbeet, M.; Canters, G. W. Appl. Phys. Lett. 2003, 82, 472. (8) Craighead, H. G.; Montemagno, C. D. Science 2000, 290, 1555. (9) Katchalski-Katzir, E. Trends Biotechnol. 1993, 11, 471. (10) Sheldon, R. A. AdV. Synth. Catal. 2007, 349, 1289. (11) Wang, P. Curr. Opin. Biotechnol. 2006, 17, 574. (12) Choi, E. J.; Foster, M. D.; Daly, S.; Tilton, R.; Przybycien, T.; Majkrzak, C. F.; Witte, P.; Menzel, H. Langmuir 2003, 19, 5464. (13) Minton, A. P. Biophys. J. 1999, 76, 176. (14) Zhou, D. J.; Wang, L. B.; Rayment, T.; Abell, C. Langmuir 2003, 19, 10557. (15) Pallandre, A.; De Meersman, B.; Blondeau, F.; Nysten, B.; Jonas, A. M. J. Am. Chem. Soc. 2005, 127, 4320. (16) Davies, D. R.; Cohen, G. H. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 7. (17) Berquand, A.; Xia, N.; Castner, D. G.; Clare, B. H.; Abbott, N. L.; Dupres, V.; Adianensen, Y.; Dufrene, Y. Langmuir 2005, 21, 5517. (18) Wadu-Mesthrige, K.; Xu, S.; Amro, N. A.; Liu, G. Y. Langmuir 1999, 15, 8580. (19) Buijs, J.; Speidel, M.; Oscarsson, S. J. Colloid Interface Sci. 2000, 226, 237. (20) Leisten, F.; Wiechmann, M.; Enders, O.; Kolb, H. A. J. Colloid Interface Sci. 2006, 298, 508.

level. Lysozyme is a well-studied protein with known structure and function.16-29 A lysozyme molecule is about 3 × 3 × 5 nm3.21 Active lysozyme had been previously immobilized onto a carboxylic acid terminated chemical pattern.16,22 The lysozyme molecules formed a tightly packed monolayer on the chemical pattern. Since the adsorption chemistry, pattern geometry, and molecular level adsorption structures of immobilized lysozyme have been well studied, we chose the lysozyme pattern as a model system to investigate the protein pattern/antibody binding interaction through direct atomic force microscope (AFM) visualization.30-33 We find that polyclonal anti-lysozyme antibodies preferentially bind to lysozyme on the pattern edge.

II. Experimental Section Our experimental procedure is illustrated in Schemes 1 and 2. The carboxylic acid terminated chemical patterns on the octadecyltrichlorosilane (OTS) monolayer film were fabricated to direct the assembly of lysozyme. The protein pattern followed the shape and dimension of the chemical pattern. Carboxylic acid terminated chemical patterns were created with AFM probe oxidation lithography.34,35 Route 1 (Scheme 1) led to a carboxylic acid terminated chemical pattern (OTSox) 10 Å lower than the OTS background,36 whereas the carboxylic acid terminated chemical pattern fabricated via route 2 (OTSoxUTSox, Scheme 2) protruded 14 Å above the OTS monolayer. (21) Sauter, C.; Ota´lora, F.; Gavira, J.-A.; Vidal, O.; Giege´, R.; Garcı´a-Ruiz, J. M. Acta Crystallogr., Sect. D: Biol. Crystallogr. 2001, 57, 1119. (22) Cai, Y.; Ocko, B. M. Langmuir 2005, 21, 9274. (23) Yamaguchi, T.; Komeda, T. Jpn. J. Appl. Phys. 2006, 45, 2349. (24) Haggerty, L.; Lenhoff, A. M. Biophys. J. 1993, 64, 886. (25) Daly, S. M.; Przybycien, T. M.; Tilton, R. D. Langmuir 2003, 19, 3848. (26) Pasche, S.; Vo¨ro¨s, J.; Griesser, H. J.; Spencer, N. D.; Textor, M. J. Phys. Chem. B 2005, 109, 17545. (27) Fang, W.; Zhu, S. P.; Ishiharar, K.; Brash, J. L. Langmuir 2005, 21, 5980. (28) Koehler, J. A.; Ulbricht, M.; Belfort, G. Langmuir 1997, 13, 4162. (29) Kim, D. T.; Blanch, H. W.; Radke, C. J. Langmuir 2002, 18, 5841. (30) Browning-Kelley, M. E.; Wadu-Mesthrige, K.; Hari, V.; Liu, G. Y. Langmuir 1997, 13, 343. (31) Cullen, D. C.; Lowe, C. R. J. Colloid Interface Sci. 1994, 166, 102. (32) Xu, H.; Grant, C.; Lu, J. R.; Williams, D. E.; Penfold, C. Langmuir 2006, 22, 6313. (33) Li, L. Y.; Chen, S. F.; Oh, S. J.; Jiang, S. Y. Anal. Chem. 2002, 74, 6017. (34) Maoz, R.; Cohen, S. R.; Sagiv, J. AdV. Mater. 1999, 11, 55. (35) Maoz, R.; Frydman, E.; Cohen, S. R.; Sagiv, J. AdV. Mater. 2000, 12, 725. (36) Cai, Y. Langmuir 2008, 24, 337.

10.1021/la801020b CCC: $40.75  2008 American Chemical Society Published on Web 08/13/2008

Boundary Molecules in a Lysozyme Pattern Scheme 1. Route 1, OTSoxa

a OTS (1a) is oxidized to COOH-terminated OTSox (1b, pink) under a conducting AFM probe. Next, lysozyme molecules (green) absorb on the OTSox, forming a lysozyme pattern (1c, otsox-lysozyme). The polyclonal antibodies (blue) bind to the edge of the lysozyme pattern first (1d). When the incubation time is sufficiently long, antibodies cover the whole lysozyme pattern (1e). (Not drawn to the actual ratio. The antibody binding structures in 1d and e are for illustration purposes showing they bind to the lysozyme. They do not mean that antibodies actually bind to lysozyme in this specific orientation).

Scheme 2. Route 2, OTSoxUTSoxa

a OTS (2a) is oxidized to OTSox (2b, pink) through the scanning probe oxidation. Next, the OTSox pattern is reacted with 10-undecenyltrichlorosilane (UTS), to form a double bond terminated OTSoxUTS pattern. The OTSoxUTS pattern is then oxidized to a COOH-terminated OTSoxUTSox pattern (2c). Next, lysozyme is immobilized on the OTSoxUTSox pattern. The immobilized lysozyme on the OTSoxUTSox template (2d, OTSoxUTSox-lysozyme) is then reacted with antibodies for 15 min (2e).

Instrumentation. The chemical pattern fabrication and characterization were performed with the Agilent PicoPlus AFM in an environmental chamber. The patterns were characterized in the tapping mode with MikroMasch NSC-14 tips, which have a typical natural frequency of 150 kHz and a force constant of 5 N/m. All AFM characterizations were performed immediately after the samples were ready. Imaging was performed in an environmental chamber with 70% relative humidity (RH). The OTS film thickness and quality were examined with an Angstrom Advance PhE 101 ellipsometer and a Varian Excalibur 3100 Fourier transformed infrared spectrometer equipped with a semiconductor cooled DTGS detector. The IR spectra of OTS films on silicon wafers were taken at 4 cm-1 resolution with 4096 scans. The infrared spectra of the protein patterns were acquired by using a Varian UMA 600 IR microscope equipped with a liquid nitrogen cooled mercury cadmium telluride (MCT) narrow band detector in the reflection mode. Preparation of Octadecyltrichlorosilane (OTS) Self-Assembled Monolayers. The preparation of the OTS films followed the established procedure, which has been described in detail elsewhere.36 Contact mode AFM analysis of the OTS film indicated that the root mean square (rms) roughness of the OTS film was smaller than 5 Å. The OTS film thickness was 26 ( 3Å, measured with the ellipsometer. The Brewster angle infrared spectra of the OTS film were comparable to published results.37 Fabrication of Carboxylic Acid Terminated Patterns (Route 1). The OTSox pattern was fabricated on self-assembled monolayers of OTS by electrochemical writing under a conductive AFM probe (37) Maoz, R.; Sagiv, J.; Degenhardt, D.; Miihwald, H.; Quint, P. Supramol. Sci. 1995, 2, 9.

Langmuir, Vol. 24, No. 18, 2008 10335 in the contact mode in a humid environment.35,36 A Mikromasch type CSC-17 platinum-titanium coated conductive AFM tip was used in the pattern fabrication process. The pattern was designed using the Agilent PicoLith software. A 5-10 V bias was applied to the silicon, and as a result the water between the tip and the sample was electrolyzed, generating oxygen radicals, which oxidized the methyl-terminated OTS film to form carboxylic acid terminated patterns under the AFM tip.35,36 A moving tip created the line patterns, while a stationary tip created the disc shaped patterns. The design of the PiciLith software determines that the AFM tip always halts briefly before changing direction during the pattern writing process. As a result, the joints between two lines always appear as round discs. Fabrication of Carboxylic Acid Terminated Patterns (Route 2). An OTSox pattern was first fabricated. Subsequently, the pattern was soaked in a 5 mM undecenyltrichlorosilane (UTS) toluene solution overnight. The UTS silane self-assembled on the hydrophilic OTSox pattern, forming a second layer, which was double bond terminated and 14 Å above the OTS background (Figure S4 in the Supporting Information). Next, the sample with the bilayer pattern (UTS on the OTSox, OTSoxUTS) was incubated in the oxidizing solution (0.01 M NaIO4, 5 × 10-4 M KMnO4, in 0.05 M Na2CO3 buffer) for 10 h at 40 °C. The sample was then rinsed with deionized water and followed by 1% hydrazine for 1 min. Finally, the sample was dried in a stream of nitrogen. Since the pattern size is in the nanometer range, it is difficult to monitor the reaction in real time with spectroscopy methods. However, we prepared a UTS monolayer on a 2.5 cm × 2.5 cm silicon wafer. The UTS coated wafer was treated using the same procedure as the oxidation process of the OTSoxUTS pattern. Figure S1 in the Supporting Information shows the Brewster angle IR spectra of the UTS wafer before and after oxidization. After oxidation, the 3080 and 1645 cm-1 peaks of the vinyl group disappeared and a new peak at 1710 cm-1 appeared which corresponded to the carboxylic acid group. Deposition of the Lysozyme onto the Chemical Pattern. The lysozyme (hen egg white) was dissolved in a 25 mM pH 7 HEPES buffer to a final concentration of 4 µg/mL. Samples bearing the chemical pattern were dipped into the solution. Next, the sample was rinsed with deionized water and gently wiped with the “KimWipe” paper to remove the nonspecifically adsorbed protein. Deposition of the Anti-Lysozyme Antibody onto the Protein Pattern. One drop (33 µL) of 100 or 200 ng/mL polyclonal antibody (anti-hen egg white, rabbit, Rockland Immunochemicals, Inc.) solution was applied to cover the pattern on the surface. When the set time was reached, the antibody solution was removed by aspiration, followed by three rinses with 100 µL of 5 mM pH 7 HEPES buffer to remove most of the remaining nonspecifically absorbed antibodies from the surface. Finally, the surface was blown dry with a stream of nitrogen. Infrared Spectra of Lysozyme Patterns. Lysozyme patterns smaller than 10 µm would not be detectable by the IR microscope. A larger lysozyme pattern is needed for IR measurement. We used AFM to fabricate an 8 × 8 array of lysozyme discs in a 50 × 50 µm2 area. Each disc in the array is 3 µm in diameter. Sixteen such arrays were fabricated as a 4 × 4 matrix inside a 200 × 200 µm2 region, resulting in 1024 lysozyme discs in total. The silicon wafer we used is polished on one side. The pattern was fabricated on the polished side. On the rough side, we applied a trace amount of Vaseline to eliminate the IR interference from two surfaces of the sample. To locate the lysozyme pattern, we made a scratch mark on the wafer before fabricating the protein pattern. Using the optical microscope attached on the IR microscope and the marks on the wafer as a reference, we located the invisible pattern and obtained the IR spectra over the 200 × 200 µm2 patterned region. To generate a standard IR spectrum for the lysozyme, one drop (5 µL) of 4 mg/mL lysozyme solution was directly applied onto the silicon wafer. After drying, the lysozyme formed a thick film on the surface, which appeared as a dark stain through the optical microscope. We acquired the IR spectrum of the thick film and used it as the standard IR spectrum of the lysozyme.

10336 Langmuir, Vol. 24, No. 18, 2008

Figure 1. Representative lysozyme pattern on the OTSox template (topography image, tapping mode, 14 µm × 14 µm, the scanning rate is 1 line/s). The histogram (inset) shows that the lysozyme pattern is 18 Å above the OTS background.

Control Experiment for Antibody Binding. One drop (33 µL) of 100 ng/mL anti-polyhistidine antibody (Sigma) solution was applied to the lysozyme pattern. One drop (33 µL) of 100 ng/mL anti-lysozyme antibody (Rockland Immunochemicals, Inc.) solution was applied to another sample with the lysozyme pattern. After rinsing, the two samples were characterized by using AFM in the tapping mode.

III. Results and Discussion Lysozyme Patterns Immobilized on the Carboxylic Acid Terminated Chemical Pattern. The lysozyme molecules were immobilized on the OTSox patterns by the attractive Coulomb force. Since the lysozyme has an isoelectric point of 11,28 in a pH 7 buffer, the lysozyme molecule carries positive charges, whereas the carboxylic acid terminated chemical pattern carries negative charges. So, upon dipping the sample with the chemical pattern into the lysozyme solution at pH 7, the attractive Coulomb force drove the lysozyme molecule to selectively adsorb onto the carboxylic acid terminated chemical pattern, forming a lysozyme pattern following the underneath chemical pattern. The AFM image in Figure 1 shows that the lysozyme pattern immobilized on OTSox (OTSox-lysozyme) is a monolayer. The histogram in the inset shows that the lysozyme pattern has an apparent height of 18 Å. Since the OTSox chemical template is 10 Å below the OTS surface,36 the height of the lysozyme is 28 Å, consistent with previously reported results.22,29 Similarly, the lysozyme pattern immobilized on OTSoxUTSox (route 2, OTSoxUTSox-lysozyme) is a 32 Å high monolayer. Although the lysozyme molecules in the pattern do not form a 2-D crystal, they appear as a continuous terrace as revealed by the 5 nm resolution AFM images (Figure S2 in the Supporting Information). Defect spots that expose the underneath OTSox are rare in the lysozyme pattern. Most of the lysozyme molecules pack tightly and touch each other. The lateral motions of the lysozyme molecules are highly restrained. Previously, under similar deposition conditions, lysozyme molecules were found to adsorb on the Au(111) and graphite surfaces. Molecular resolution scanning tunneling microscopy (STM) revealed that these lysozyme molecules formed a tightly packed monolayer, which is consistent with our AFM results.23,24

Gao and Cai

Figure 2b shows the infrared spectra of the lysozyme monolayer immobilized on the chemical template. Compared with the infrared spectra of the lysozyme thick film (Figure 2a), the amide I, amide II, amide III, and N-H stretching bands are all visible, although they are very weak due to the monolayer nature of the pattern. Our infrared spectra of the patterned region are similar to the spectra of absorbed lysozyme from other groups.38-42 “Boundary Binding” Effect of the Lysozyme Pattern. The OTSox-lysozyme pattern was incubated with the polyclonal anti-lysozyme antibody in a pH 7 HEPES buffer. At 100 ng/mL concentration and 15 min incubation time, the antibodies preferentially bound on the edge of the lysozyme pattern while only a few antibodies bound to the lysozyme molecules in the pattern interior. Figure 3 shows the lysozyme pattern immediately after the antibody incubation. After incubation with the antibody, the edge of the lysozyme pattern grew an additional 28 Å in height, which corresponds to the height of the antibody.31-33 After incubating the lysozyme pattern for 8 h with a 200 ng/mL antibody solution, the antibody covered the whole lysozyme pattern. Figure 2c shows the infrared spectrum of the antibody covered lysozyme pattern, where a broad peak from 1100-1900 cm-1 appears and buries all weak peaks of the amides. Our control experiments showed that the rinsing process after the antibody deposition did not affect the adsorption of the lysozyme or the antibodies bound to the lysozyme. We rinsed a lysozyme pattern that had been incubated for 8 h with a 200 ng/mL antibody solution. The pattern was fully covered by the antibody. We imaged the sample before and after the rinsing. We found that the rinsing only removed the nonspecifically adsorbed antibodies on the OTS. There was no antibody leaching on the pattern during rinsing. To demonstrate the observed binding is a biorecognition event, anti-polyhistidine antibodies were incubated with the lysozyme pattern following the same procedure to serve as a control test (Figure 4). While many molecules of anti-lysozyme antibodies bind preferentially to the edge of the protein pattern, the antipolyhistidine antibodies do not show “edge binding”. Therefore, the observed “boundary binding” is a biospecific binding rather than nonspecific adsorption of antibodies on the step edges. Our observation indicates that (i) the immobilized lysozyme pattern maintains the antibody binding activity and (ii) the antibodies bind to the boundary lysozyme molecules of a protein pattern before they bind to the interior lysozyme molecules. We compared antibody binding to lysozyme molecules in a tightly packed, full monolayer lysozyme pattern and in a loosely packed, submonolayer pattern. Figure 5 shows results obtained for both types of lysozyme patterns.43 The lysozyme film on the disc shaped pattern forms a tightly packed, full monolayer. The line pattern was created using a fast moving tip during the electrochemical oxidation of the OTS film, which led to incomplete oxidation in the pattern, whereas the disc pattern was created using a stationary tip, which brought the complete oxidation of the OTS film. Thereby, the negative charge density over the line is lower than that of the disc pattern. As a result, after the lysozyme deposition, the lysozyme has only a sub(38) Yang, T.; Li, Z.; Wang, L.; Guo, C. L.; Sun, Y. J. Langmuir 2007, 23, 10533. (39) Rangnekar, A.; Sarma, T. K.; Singh, A. K.; Deka, J.; Ramesh, A.; Chattopadhyay, A. Langmuir 2007, 23, 5700. (40) Basar, N.; Uzun, L.; Gu¨ner, A.; Denzili, A. J. Appl. Polym. Sci. 2008, 108, 3454. (41) Liltorp, K.; Mare´chal, Y. Biopolymers 2005, 79, 185. (42) Hadden, J. M.; Chapman, D.; Lee, D. C. Biochim. Biophys. Acta 1995, 1248, 115. (43) Processed with the WSxM. Horcas, I.; Fernandez, R.; Gormez-Rodrigues, J. M.; Colchero, J.; Gomez-Herrero, J.; Baro, A. M. ReV. Sci. Instrum. 2007, 78, 013705.

Boundary Molecules in a Lysozyme Pattern

Langmuir, Vol. 24, No. 18, 2008 10337

Figure 2. Infrared spectra of lysozyme and lysozyme-antibody complex: (a) infrared spectrum of a thick lysozyme film on the surface, (b) infrared spectrum of the lysozyme monolayer immobilized on the chemical template, and (c) spectrum of the same pattern after incubation with the antibody.

Figure 3. Polyclonal antibodies preferentially bind to the edge of the OTSox-lysozyme pattern after 15 min incubation in a 100 ng/mL antibody solution. (a) Topography image (tapping mode, 2.3 µm × 2.3 µm, the scanning rate is 0.8 line/s.). (b) Height cross-sectional profile corresponding to the green line in (a). The profile shows that the lysozyme pattern is 18 Å above the OTS. The edge of the lysozyme pattern increases by an additional 28 Å after incubating in the antibody solution. (c) Structure model of the antibody bound on the lysozyme pattern.

Figure 4. Control test of the antibody binding. Representative AFM tapping mode topography images; both (a) and (b) are 917 nm × 917 nm in size and 8.0 nm in height scale. A concentration of 100 ng/mL anti-lysozyme polyclonal antibody (a) or anti-polyhistidine antibody (b) was deposited on two lysozyme patterns with the same conditions. Most of the anti-lysozyme antibodies preferentially bind to the edge of the disc shaped lysozyme pattern (white spots). Just a few anti-polyhistidine antibodies are observed on the lysozyme pattern, with only one absorbed close to the edge.

monolayer coverage on the line patterns. Figure 5 shows the pattern after a 5 min incubation with a 100 ng/mL antibody solution. The antibodies (white dots in the images) bound to the loosely packed, submonolayer lysozyme in the line pattern before the tightly packed monolayer ones in the disc pattern. The “preferential edge binding” did not occur for the line pattern. Most observed antibodies bound to the interior molecules of the

Figure 5. Antibody binding to the loosely packed, submonolayer lysozyme pattern (lines) and the tightly packed, full monolayer lysozyme pattern (discs).

line pattern. At this shortened incubation time, no binding was observed for the disc, indicating the loosely packed lysozyme

10338 Langmuir, Vol. 24, No. 18, 2008

molecules in the line pattern have better accessibility and improved flexibility than the edge lysozyme molecules in the tightly packed disc pattern. We speculate that the difference in the spatial accessibility and the structural flexibility between the lysozyme molecules on the edge and in the pattern center cause the observed difference in binding preference. In a tightly packed lysozyme pattern, most of the lysozyme molecules are embraced by neighboring lysozyme molecules. Both the neighboring lysozyme molecules and the underneath supporting surface block the epitopes on the lysozyme from external access. Also, the relaxation of the lysozyme molecule in the center is suppressed by its tightly packed neighbors, whereas the edge proteins are more flexible.15 Therefore, the antibody molecules have more chances to access the lysozyme on the edge and tend to bind to the edge first. When the edge sites have been occupied and there are still free antibodies in the solution, the antibodies can also bind to the interior lysozyme of the pattern. In the presence of both the disordered submonolayer lysozyme pattern and the tightly packed, full monolayer pattern, the antibodies bind to the loosely packed, submonolayer pattern before they bind to the lysozyme molecules in the tightly packed pattern, as demonstrated in Figure 5. This effect further confirms our speculation. Since the lysozyme molecules in the submonolayer film are not in contact with other molecules, they are better relaxed and easier to access than the lysozyme molecules in the tightly packed lysozyme pattern. Topographic Effect of the Lysozyme Pattern on Antibody Binding. The lysozyme on the OTSox chemical pattern is halfburied in the 10 Å deep OTSox channels, with an apparent height of only 18 Å over the OTS, whereas the lysozyme on the OTSoxUTSox pattern is immobilized on a 14 Å high terrace with an apparent height of 46 Å above the OTS. To investigate the topographic effect of the lysozyme pattern toward the antibody binding, we deposited the antibodies on the OTSoxUTSoxlysozyme pattern following the same procedure as described for the OTSox-lysozyme pattern. Figure 6 shows the OTSoxUTSox-lysozyme pattern after 15 min incubation with a 100ng/ mL antibody solution. At first glance, the pattern appears the same before and after the antibody incubation. However, by analyzing the topographic image (Figure 6a) together with the corresponding phase image (Figure 6b), we find that the antibodies also preferentially bound to the edge. Furthermore, these antibodies bound to the edge in a configuration different from the antibody binding configuration in the OTSox-lysozyme pattern. We first identify the positions of antibodies in the pattern through the phase signal. The phase channel detects the chemical properties of a surface. Although both the lysozyme and antibody are proteins, they have quite different contrasts in the phase image. The antibody has a higher phase contrast than both the lysozyme and OTS, and it appears as a white spot in the phase image.44 The topographic change may also interfere with the phase signal. However, the phase signal jumpings due to the topographic change have unique characteristics and can be distinguished from the phase signal changes from the surface chemistry. Figure 7 shows a lysozyme pattern immobilized on the OTSoxUTSox template. The lysozyme pattern appears as a terrace 46 Å above the OTS. When the tip moved across a step edge, the sudden topography change caused “jumping” of the phase signal. However, in such circumstances, the direction of jumping is associated with the scanning direction. In Figure 7d, (44) Please see the Supporting Information, section 3 for the analysis and discussion about the phase contrast of the antibody.

Gao and Cai

Figure 6. Antibodies preferentially bind to the OTSoxUTSox-lysozyme pattern from the side. (a) Topography image, tapping mode, 1.4 µm × 1.4 µm, 0.8 line/s. Inset: The lysozyme disc is rendered in an exaggerated color scale to demonstrate the 31 Å high antibody on the edge. Due to the minute height difference between the lysozyme pattern and the antibody, the common gradient scale is unable to show the antibody in the topographic image. (b) Phase image. A white rim around the pattern edge reveals the presence of antibodies on the edge. (c) Height (purple line) and phase (blue line) cross-sectional profiles. The phase signal has two peaks over the pattern edge, where the topographic data show that the height of the white rim is 31 Å. This height matches the size of the antibody. In the phase profile, there are also few peaks between the two “edge peaks”. They correspond to the few antibodies bound to the lysozyme pattern from the top. (d) Structure model of the side-bound antibody on the lysozyme pattern.

Figure 7. Lysozyme pattern immobilized on the OTSoxUTSox template. The lysozyme is 46 Å above the OTS. The phase signal peaks from the topographic jumping are opposite for the trace and retrace scans (3.8 µm × 3.8 µm, tapping mode). (a) Phase image, trace (from left to right, 1.5 V phase scale). (b) Phase image, retrace (from right to left, 1.5 V phase scale). (c) Topographic image, 15 nm height scale. (d) Phase signal cross-sectional profiles corresponding to the red (trace) and green (retrace) lines across the lysozyme terrace.

the green curve corresponds to the green line in the trace (left to right scan) phase image (Figure 7a). The left (rising) edge and the right (falling) edge appear as a positive peak and a negative

Boundary Molecules in a Lysozyme Pattern

peak, respectively, in the phase cross-sectional plot. In the right to left scan phase image (Figure 7b), the left (falling) edge and the right (rising) edge appear as a negative peak and a positive peak (red curve in Figure 7d), respectively. Another characteristic associated with the topographic originated phase change is that the phase changes in the trace scan line and the retrace scan line have the opposite directions. The green and red curves in Figure 7d are the trace and retrace phase cross-sectional profiles corresponding the same topographic position. They have opposite peaks over the same topographic edge. Therefore, based on these two unique characteristics, we can distinguish the phase signal changes originating from the topography and the phase signal changes originating from the surface chemistry. In the phase channel image of the lysozyme pattern on the OTSoxUTSox template (Figure 6b) after the antibody incubation, a white rim surrounds the lysozyme disc pattern in the phase channel. Compared with the representative plots of the phase channel over the lysozyme pattern before the antibody incubation (Figure 7d) and after the antibody incubation (Figure 6c, blue line), we find they are different. After the antibody incubation, at the rising edge and the falling edge, the phase signal shows two positive peaks. In addition, the trace and retrace phase signals are identical. From the discussion above, we conclude that these peaks in the phase channel are not from topographic interference. They indicate the presence of the antibody surrounding the lysozyme pattern after the antibody incubation. Therefore, the white rim around the lysozyme pattern is assigned to the bound antibodies. Figure 6 shows that, under the same antibody exposure level, the antibodies also preferentially bind to the edge lysozyme molecules in the OTSoxUTSox-lysozyme pattern. However, the antibodies preferentially bound to the edge lysozyme molecules from the top to the OTSox-lysozyme pattern, while they bound from the side to the OTSoxUTSoxlysozyme pattern. In Figure 6c, we plot the height cross-sectional profile and the phase cross-sectional profile together. At the pattern edges, the phase signal has two peaks, which correspond to the antibody positions. In the same positions of the height crosssectional profile (along the green dashed lines in Figure 6c), we find there are two small steps on the edge. Based on the phase signal, they should be the antibodies. Even though these antibodies

Langmuir, Vol. 24, No. 18, 2008 10339

are difficult to visualize directly, the height cross-sectional plot of these steps reveals that their heights are 31 Å. This value matches the reported height of the antibody.31-33 If the antibodies bind to the edge lysozyme from the top, the antibody would be higher than the protein pattern, which is never observed for the OTSoxUTSox-lysozyme pattern. Since the lysozyme is 30 Å in height and sits on the 14 Å high OTSoxUTSox template, the observed 31 Å height of these antibodies on the protein pattern edge indicates that the antibodies bind to the protein from the side (Figure 6d). In the phase profile, there are also few peaks between the two “edge peaks”. They correspond to the few antibodies bound to the lysozyme pattern from the top. The AFM images reveal that the fully extruding lysozyme pattern favors the side binding. Most of the antibodies bind to the pattern from the side; only a few bind to the lysozyme pattern from the top. In contrast, the half-buried lysozyme pattern favors the top binding. Such a difference in the binding direction indicates that the “template embedding” approach in protein immobilization significantly affects the protein binding toward the polyclonal antibody.

IV. Conclusion We fabricated a tightly packed lysozyme pattern on a prefabricated chemical pattern. We find that the antibodies preferentially bind to the edge lysozyme molecules, since they have better spatial accessibility and flexibility. If the coverage of an immobilized protein remains the same, a protein nanoarray has more protein molecules on the edge than a protein microarray. In turn, a protein microarray has more protein molecules on the edge than a protein film. Therefore, reducing the protein pattern feature to the nanoscale will improve the overall binding of immobilized proteins toward antibodies. Acknowledgment. This research is supported by the University of Kentucky faculty start-up grant. Supporting Information Available: Detailed AFM image analysis of the protein patterns and the antibodies on the surface. This material is available free of charge via the Internet at http://pubs.acs.org. LA801020B