Orientations of Nematic Liquid Crystals on Surfaces Presenting

Presenting Controlled Densities of Peptides: Amplification of Protein-Peptide Binding Events. Brian H. Clare and Nicholas L. Abbott*. Department of Ch...
0 downloads 0 Views 950KB Size
Download PDF
Langmuir 2005, 21, 6451-6461

6451

Orientations of Nematic Liquid Crystals on Surfaces Presenting Controlled Densities of Peptides: Amplification of Protein-Peptide Binding Events Brian H. Clare and Nicholas L. Abbott* Department of Chemical and Biological Engineering, University of Wisconsin-Madison, 1415 Engineering Drive, Madison, Wisconsin 53706 Received February 4, 2005. In Final Form: April 16, 2005 We report a study of the orientations of nematic liquid crystals (LCs) in contact with peptide-modified, oligoethylene glycol-containing, self-assembled monolayers (SAMs). The SAMs were formed on gold films that were prepared by physical vapor deposition at an oblique angle of incidence. Two peptides were investigated: the optimized substrate for the Src protein kinase (IYGEFKKKC) and the synthetic equivalent of that peptide after kinase modification (IpYGEFKKKC). Polarization modulation-infrared reflectance absorbance spectroscopy (PM-IRRAS) was used to characterize the relative areal densities and orientations of these peptides at the interface. We conclude that the presence/absence of a phosphate group can influence the maximum packing density of immobilized peptide. We evaluated the orientations of the nematic liquid crystal 5CB in contact with these peptide-modified surfaces by using polarized microscopy. The time required for the nematic phase of 5CB to exhibit long-range orientational ordering (uniform alignment) was found to increase with increasing areal densities of immobilized peptide. We also found that the specific binding event between anti-phosphotyrosine IgG and the surface-immobilized phosphopeptide leads to an increase in the time required for the liquid crystal to achieve uniform anchoring (exceeding the experimentally accessible time scales). These results, when combined, suggest that the areal density and size of biomolecules at an interface can influence the time required for liquid crystals in contact with nanostructured surfaces to exhibit long-range orientational order. Finally, we illustrate the potential utility of this system by demonstrating that liquid crystals can be used to amplify and report protein binding events occurring on a spatially resolved peptide array.

Introduction Peptide-modified interfaces are emerging as useful tools for the study and high-throughput screening of enzymatic activity,1-4 control of cellular behavior,5,6 and the capture of protein analytes from complex mixtures.7,8 In comparison to the analogous protein-chips,9 peptides at interfaces (1) are synthetically accessible using solid-phase methods and can be purified quickly; (2) exhibit higher stability toward heat and pH variation; and (3) generally present only one functional motif, whereas proteins often contain multiple recognition motifs.10 Several materials have previously been explored for the attachment of peptides to interfaces, including glass slides,2,6,11 titanium oxide,12 polymers,10 and metals such as gold.13 The use of self* Corresponding author. Phone: (608) 265-5278. Fax: (608) 2625434. E-mail: [email protected]. (1) Salisbury, C. M.; Maly, D.; Ellman, J. A. J. Am. Chem. Soc. 2002, 124, 14868. (2) Falsey, J. R.; Renil, M.; Park, S.; Li, S.; Lam, K. S. Bioconjugate Chem. 2001, 12, 346. (3) Houseman, B. T.; Mrksich, M. Trends Biotechnol. 2002, 20, 279. (4) Schutkowksi, M.; Reimer, U.; Panse, S.; Dong, L.; Lizcano, J. M.; Alessi, D. R.; Schneider-Mergener, J. Angew. Chem., Int. Ed. 2004, 43, 2671. (5) Shin, H.; Jo, S.; Mikos, A. G. Biomaterials 2003, 24, 4353. (6) Matsuzawa, M.; Umemura, K.; Beyer, D.; Sugioka, K.; Knoll, W. Thin Solid Films 1997, 305, 74. (7) Schulze, W. X.; Mann, M. J. Biol. Chem. 2004, 279, 10756. (8) Duburcq, X.; Olivier, C.; Malingue, F.; Desmet, R.; Bouzidi, A.; Zhou, F.; Auriault, C.; Gras-Masse, H.; Melnyk, O. Bioconjugate Chem. 2004, 15, 307. (9) MacBeath, G.; Schreiber, S. L. Science 2000, 289, 1760. (10) Hersel, U.; Dahmen, C.; Kessler, H. Biomaterials 2003, 24, 4385. (11) Lesaicherre, M.-L.; Uttamchandani, M.; Chen, G. Y. J.; Yao, S. Q. Bioorg. Med. Chem. Lett. 2002, 12, 2079. (12) Xiao, S.-J.; Textor, M.; Spencer, N. D. Langmuir 1998, 14, 5507. (13) Petoral, R. M.; Herland, A.; Broo, K.; Uvdal, K. Langmuir 2003, 19, 10304.

assembled monolayers (SAMs) of organothiol compounds on gold to fabricate these chips14-17 permits the terminal functionality of the thiols to be chosen either for the attachment of biomolecules or tailored to resist the nonspecific adsorption of proteins and cells at interfaces.18-21 Additionally, the areal density of a desired terminal functionality can be controlled by using twocomponent mixed SAMs.22 This paper describes an investigation of the orientations of liquid crystals in contact with surfaces that present covalently immobilized peptides. This investigation is founded on the observation that the optical properties and surface sensitivity of nematic liquid crystals can form the basis of label-free methods for reporting chemical23,24 and biomolecular25,26 events occurring at interfaces. In past studies, the binding of a protein to a nanostructured (14) Houseman, B. T.; Huh, J. H.; Kron, S. J.; Mrksich, M. Nat. Biotechnol. 2002, 20, 270. (15) Houseman, B. T.; Gawalt, E. S.; Mrksich, M. Langmuir 2003, 19, 1522. (16) Wegner, G. J.; Lee, H. J.; Corn, R. M. Anal. Chem. 2002, 74, 5161. (17) Rigler, P.; Ulrich, W.-P.; Hoffman, P.; Mayer, M.; Vogel, H. ChemPhysChem 2003, 4, 268. (18) Lahiri, J.; Isaacs, L.; Tien, J.; Whitesides, G. M. Anal. Chem. 1999, 71, 777. (19) Prime, K. L.; Whitesides, G. M. Science 1991, 252, 1164. (20) Luk, Y.-Y.; Kato, M.; Mrksich, M. Langmuir 2000, 16, 9604. (21) Kane, R. S.; Deschatelets, P.; Whitesides, G. M. Langmuir 2003, 18, 2388. (22) Spinke, J.; Liley, M.; Schmitt, F.-J.; Guder, H.-J.; Angermaier, L.; Knoll, W. J. Chem. Phys. 1993, 99, 7012. (23) Shah, R. R.; Abbott, N. L. Science 2001, 293, 1296. (24) Shah, R. R.; Abbott, N. L. J. Am. Chem. Soc. 1999, 121, 11300. (25) Gupta, V. K.; Skaife, J. J.; Dubrovsky, T. B.; Abbott, N. L. Science 1998, 279, 2077. (26) Brake, J. M.; Daschner, M. K.; Luk, Y.-Y.; Abbott, N. L. Science 2003, 302, 2094.

10.1021/la050336s CCC: $30.25 © 2005 American Chemical Society Published on Web 06/10/2005

6452

Langmuir, Vol. 21, No. 14, 2005

Figure 1. Use of nematic liquid crystal 5CB (A) to report biomolecule binding events at interfaces. (B) The binding of protein changes the ordering of the liquid crystal near the interface. (C) Images of samples (before and after protein binding) when viewed using a polarizing microscope.

surface was shown to cause measurable changes in the orientations and optical appearances of liquid crystals (Figure 1). As peptide-modified interfaces find broad application in the study of enzymes, protein binding events and control of cellular behavior, we sought to design these surfaces such that they can be used in combination with liquid crystals. Here, we report a strategy to modify interfaces with peptides such that these materials can be used to amplify and report a range of peptide-mediated interfacial phenomena. Nematic liquid crystals are materials with mobilities characteristic of liquids yet are capable of organizing over distances of hundreds of micrometers.27 Past theoretical and experimental studies have established that the orientations of liquid crystals near an interface to a confining medium are dictated by the chemical and topographical structure of that interface.28,29 Methods to control the orientations of liquid crystals include using surfaces of solids with anisotropic topography prepared by oblique deposition of metals30,31 or lithographic processes.32 In addition, the orientations of liquid crystals can be affected by the chemical functional groups presented at interfaces via, for example, hydrogen bonding,33 presence of an electrical double layer,34 and metal-ligand interactions.23 The macroscopic orientation of the terminal groups in a SAM formed on gold films prepared by oblique deposition35 can influence the directionality of these interactions and lead to preferred azimuthal orientations of liquid crystals near the interface.36 This so-called (27) de Gennes, P. G. The Physics of Liquid Crystals, 1st ed.; Oxford University Press: London, 1974. (28) Jerome, B. Rep. Prog. Phys. 1991, 54, 391. (29) Cognard, J. Mol. Cryst. Liq. Cryst. Suppl. 1982, 78, 1. (30) Gupta, V. K.; Abbott, N. L. Langmuir 1996, 12, 2587. (31) Janning, J. L. Appl. Phys. Lett. 1972, 21, 173. (32) Kim, S.-R.; Teixeira, A. I.; Nealey, P. F.; Wendt, A. E.; Abbott, N. L. Adv. Mater. 2002, 14, 1468. (33) Luk, Y.-Y.; Yang, K.-L.; Cadwell, K.; Abbott, N. L. Surf. Sci. 2004, 570, 43. (34) Shah, R. R.; Abbott, N. L. J. Phys. Chem. B 2001, 105, 4936. (35) Follonier, S.; Miller, W. J. W.; Abbott, N. L.; Knoesen, A. Langmuir 2003, 19, 10501. (36) Gupta, V. K.; Abbott, N. L. Phys. Rev. E 1996, 54, R4540.

Clare and Abbott

anchoring of liquid crystals by surfaces28,29 has found widespread use in the display industry and underlies the principles that are being developed for the detection of molecular and biomolecular events at interfaces: a change in the chemical or topographical structure of an interface brought about by a chemical or biological species at a surface can give rise to new orientations of liquid crystals in contact with that surface.23,25,26,37,38 As liquid crystals are birefringent, these new orientations can be visualized under simple polarized microscopy. In this paper, we report methods to manipulate the areal density of surface-immobilized peptides on nanostructured gold films so as to achieve control over the orientation of liquid crystals on these surfaces. The nanometer-scale topography of these films is introduced using physical vapor deposition of gold at an oblique angle of incidence. These films have been previously characterized using AFM to possess statistical corrugations with an amplitude of 3-5 nm and wavelength of ∼30 nm.39 On these gold films, we immobilize a peptide sequence that is a known substrate for the Src protein kinase.15 This protein kinase has broad biological importance, as it is implicated in aggressive forms of colon and breast cancer,40-42 and plays a role in the focal adhesion contact formation of migrating cells.41 We report the preparation of surfaces presenting this peptide and synthetic equivalent of this peptide after kinase modification, Src-tide and p-Src-tide, respectively (shown in Figure 2). As liquid crystals are sensitive to both the chemical functionality and the topography of an interface, care was taken to establish control of areal density and site-selectivity of the immobilized peptide. To achieve site-selectivity, we used cysteine-terminated peptides and their reactions with surface-immobilized maleimide groups.6,12,15 We achieve control over the density of presented maleimide groups (leading to controlled densities of immobilized peptides) using a method based on commercially available reagents (Scheme 1). We studied the orientations of the nematic liquid crystal, 5CB, in contact with these peptide-modified surfaces and demonstrate that specific binding of the protein anti-phosphotyrosine immunoglobulin G (IgG) to the surface-immobilized phosphorylated peptide, p-Srctide, leads to changes in the ordering of liquid crystals. As this protein-binding event is phospho-specific, we conclude that liquid crystals can be used to report the phosphorylation state of surface-immobilized peptides. Methods that report the presence of phosphorylated peptides at interfaces are useful in the study of protein kinase enzymes.2,4,14 The work presented in this paper is organized into four parts. First, we describe the preparation and spectroscopic characterization of SAMs that present well-defined areal densities of site-specifically immobilized peptides. Second, we report the orientations of the nematic liquid crystal 5CB in contact with these surfaces with the aim of identifying optimal areal densities of peptides presented on nanostructured gold surfaces that give rise to preferred orientations of 5CB. Third, we use these optimized surfaces to detect specific protein-peptide binding events, using phospho-specific IgGs. Finally, we adapt this method to (37) Skaife, J. J.; Brake, J. M.; Abbott, N. L. Langmuir 2001, 17, 5448. (38) Luk, Y.-Y.; Tingey, M. L.; Hall, D. J.; Israel, B. A.; Murphy, C. J.; Bertics, P. J.; Abbott, N. L. Langmuir 2003, 19, 1671. (39) Skaife, J. J.; Abbott, N. L. Chem. Mater. 1999, 11, 612. (40) Martin, G. S. Nat. Rev. Mol. Cell. Biol. 2001, 2, 467. (41) Frame, M. C. Biochim. Biophys. Acta 2002, 1602, 114. (42) Luttrell, D. K.; Lee, A.; Lansing, T. J.; Crosby, R. M.; Jung, K. D.; Willard, D.; Luther, M.; Rodriguez, M.; Berman, J.; Gilmer, T. M. Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 83.

Orientations of Nematic Liquid Crystals on Surfaces

Langmuir, Vol. 21, No. 14, 2005 6453

Scheme 1. Chemistry Used To Prepare Peptide-Functionalized Gold Surfacesa

a (A) SAMs formed from two-component ethanolic solutions of thiols. (B) Covalent modification of SAMs using heterobifunctional linker SSMCC and cysteine-terminated peptides.

Figure 2. Cysteine-containing peptides used in this study. (A) Optimized peptide substrate of the Src kinase protein. (B) Synthetic peptide equivalent to the phosphorylated peptide product after Src kinase modification.

image a protein-peptide binding event occurring on a simple, spatially resolved array of peptides. Materials and Methods Chemicals and Peptide Materials. All materials were used as received unless otherwise noted. Peptides were obtained from the University of Wisconsin Biotechnology Center. The identity of each peptide was confirmed by MALDI-TOF and was known to be >98% purity by analytical HPLC. Triethylene glycol-

terminated thiol (EG3) was synthesized according to previously published methods.43 The amine-terminated thiol (EG3-N) was obtained from Prochimia (Gdansk, Poland) as a hydrochloride salt. The sulfo-succinimidyl 4-(N-maleimidomethyl)cyclohexane1-carboxylate (SSMCC) linker was obtained from Pierce Biotechnology (Rockford, IL). The liquid crystal 4-cyano-4′-pentylbiphenyl (5CB) was obtained from EM Industries (New York, NY). Monoclonal anti-phosphotyrosine IgG was obtained from Sigma Aldrich. All other materials were obtained from SigmaAldrich unless noted otherwise. Gold Substrate and Preparation of Peptide-Modified SAMs. Obliquely deposited gold films (12 nm in thickness) on piranha cleaned glass slides were prepared according to procedures outlined in a prior publication.39 These anisotropic gold surfaces were prepared using a fixed deposition angle of 45° (measured with respect to the surface normal). As these gold films are semi-transparent, PM-IRRAS and ellipsometric thickness measurements were performed using reflective gold films prepared by sequentially depositing 10 nm of Ti and 200 nm of Au onto silicon wafers (Silicon Sense, Nashua, NH) at normal incidence. All gold samples were used within 1 week of preparation. The conditions used to prepare these surfaces are depicted in Scheme 1. First, SAMs were formed by immersion of the gold films in 0.10 mM ethanolic solutions of thiols EG3 and EG3-N for no less than 18 h, and then rinsed with copious amounts of water and ethanol. SSMCC heterobifunctional linker was applied to these surfaces for 45 min as a 2 mM aqueous solution in 0.1 M triethanolamine (TEA) buffer, pH 7.0 adjusted using concentrated HCl. These samples were rinsed briefly in water, and a solution of peptide (250 µM in TEA) was applied for 3 h. Next, these surfaces were rinsed 2 × 1.5 mL × 5 min TEA + 0.1% Triton-X 100 (TX). Unreacted maleimide groups on the surface were quenched using a freshly prepared 2 mM solution of 2-mercaptoethanol in phosphate buffered saline (PBS), pH 7.4, for 10 min. Samples were then rinsed with water and dried. Preparation of Peptide Arrays. SAMs composed of EG3-N and EG3 thiols were formed on films of obliquely deposited gold films as described above. The entire surface was then treated with a 2 mM solution of SSMCC in TEA. Next, ∼2.5 µL of a 250 µM solution of peptide in TEA was applied to this surface as spots (having lateral dimensions of ∼1 mm). After 3 h, the entire surface was rinsed 2 × 1.5 mL × 5 min TEA + 0.1% TX. The remaining maleimide groups were quenched by treatment of the (43) Pale-Grosdemange, C.; Simon, E. S.; Prime, K. L.; Whitesides, G. M. J. Am. Chem. Soc. 1991, 113, 12.

6454

Langmuir, Vol. 21, No. 14, 2005

entire surface with a 2 mM solution of 2-mercaptoethanol in PBS for 10 min. Finally, these samples were rinsed and dried. Protein Binding Studies Using Peptide-SAMs. Monoclonal anti-phosphotyrosine IgG was applied to the peptide surfaces for 1.5 h as a 10 µg/mL solution in PBS + 0.05% TX (approximately 67 nM when assuming MW of antibody ∼150 000 g/mol). Control samples were prepared by treating similar peptide surfaces with a 10 µg/mL solution of monoclonal anti-avidin IgG (∼67 nM) in PBS + 0.05% TX for an equal length of time. All samples were rinsed for 15 s in PBS + 0.05% Triton-X 100, then with water, and finally dried under a stream of N2. Preparation of Octyltrichlorosilane (OTS)-Treated Glass Slides. Piranha cleaned glass slides (cleaning procedures outlined in prior publication39) were immersed in a 10 mM solution of OTS in anhydrous n-heptane. After 30 min, each slide was rinsed with dichloromethane and dried under a stream of N2. Preparation of Optical Cells for Observation of 5CB Orientations. Optical cells for use in combination with polarized light microscopy were prepared to determine the orientations of liquid crystals in contact with the peptide-modified surfaces. The optical cell was fabricated by spacing two identically treated surfaces approximately 12 µm apart using thin strips of Saran wrap. To image the peptide arrays, hybrid cells were prepared using a surface on which peptides were patterned and a second surface that was OTS-treated glass. For all samples, the surfaces were held together at each end using bulldog clips and were warmed to approximately 40 °C. 5CB, heated to its isotropic phase (∼35 °C), was spontaneously drawn into each optical cell by capillary action. The optical cell was cooled to room temperature. During the cooling process, 5CB changed from its isotropic state to its nematic state. The optical appearance of the sample was observed in transmission mode using a polarized light microscope (see Image Capture). Ellipsometry. A Rudolph AutoEL ellipsometer (wavelength of 632 nm, 70° angle of incidence) was used to determine the optical thickness of the SAMs, peptides, and proteins on the surfaces of 2000 Å thick gold films. Ellipsometric constants were determined at five locations on each sample. A simple slab model was then used to interpret these constants. The slab (SAM, peptide, and protein) was assumed to have an index of refraction of 1.46. Image Capture and Luminosity Analysis. Images of the optical appearance of the liquid crystals were captured with a digital camera mounted on a polarized light microscope (BX60, Olympus). Consistent settings of both the microscope light source (aperture set at 1/2 maximum, and lamp intensity also set at 1/2 maximum) and the digital camera (2.8 f-stop, 1/650 shutter speed) allowed for the direct comparison of images taken of different samples. To quantify the luminosity of the liquid crystal in contact with the peptide arrays, each composite image was converted to a gray-scale image. The average pixel brightness of a region was calculated, assigning a completely black pixel the value of 0 and a completely white pixel receives the value of 255. PM-IRRAS. IR spectra of SAMs supported on gold films (thickness of 2000 Å) were obtained using a Nicolet Magna-IR 860 FT-IR spectrometer with photoelastic modulator (PEM-90, Hinds Instruments, Hillsboro, OR), synchronous sampling demodulator (SSD-100, GWC Technologies, Madison, WI), and a liquid N2-cooled mercury cadmium telluride (MCT) detector. All spectra were taken at an incident angle of 83° with the modulation centered at 1800 cm-1. For each sample, 500 scans were taken at a resolution of 4 cm-1. Data were collected as differential reflectance versus wavenumber, and spectra were normalized and converted to absorbance units via the method outlined in Frey et al.44 For quantitative analysis, spectra were fit to multiple Gaussian peaks using Igor Pro 4. Residuals were minimized, and the areas of each peak were determined.

Results and Discussion Part I. Preparation and Characterization of MaleimideModified Surfaces. The preparation of maleimide-modified (44) Frey, B. L.; Robert, M. C.; Weibel, S. C. Polarization-Modulation Approaches to Reflection-Absorption Spectroscopy. In Handbook of Vibrational Spectroscopy; Griffiths, P. R., Ed.; John Wiley & Sons: New York, 2002; Vol. 2; p 1042.

Clare and Abbott

SAMs is depicted in Scheme 1. A series of two-component SAMs on gold were prepared using thiols 1 and 2. The solution compositions of thiol 2 (EG3-N) used in our experiments were 1, 5, 10, 25, 50, and 100 mol %. Hereafter, we refer to these surfaces by the composition of the ethanolic solutions used to form each SAM; that is, a 5% EG3-N SAM was formed from an ethanolic solution comprised of 5 mol % EG3-N and 95 mol % EG3 thiols. Coadsorption of thiols having similar structure and length is unlikely to lead to segregation of species within the mixed SAM.45,46 These surfaces were then treated with the heterobifunctional linker SSMCC (see Materials and Methods). Polarization modulation-infrared reflectance absorbance spectroscopy (PM-IRRAS)44,47 and ellipsometry were used to (1) confirm the attachment of the maleimide group as depicted in Scheme 1B and (2) confirm control over the areal density of peptide presented on surfaces known to be largely resistant to the nonspecific adsorption of proteins.19 PM-IRRAS is a surface-sensitive analytical technique that can provide information about the quantity, type, and orientation of organic functional groups present at an interface.48 Shown in Figure 3A are the PM-IRRAS spectra obtained using mixed SAMs formed from thiols EG3 and EG3-N following treatment with SSMCC. Strong absorption bands are observed for the maleimide asymmetric (1707 cm-1) and symmetric (1745 cm-1) stretching modes. These absorption bands were previously observed by Xiao, Textor, and Spencer for peptide-modified titanium surfaces.12 The reaction of SSMCC with the SAM also generates one amide bond. We observe a band in the 1655 cm-1 region, corresponding to the amide I (CdO) stretching mode.49 The magnitudes of the absorbance peaks shown in the PM-IRRAS spectra in Figure 3A depend on the orientation and the number density of functional groups at the interface. Therefore, to make statements regarding the relative amount of SSMCC at the interface, it is necessary to determine if the orientation of the maleimide changes as a function of monolayer composition. The orientations of organic functional groups have previously been determined using infrared spectroscopy, most notably the C-H bonds present in alkanethiols chemisorbed to gold surfaces.50 When performing PM-IRRAS on gold films, only stretching modes parallel to the surface normal are observed.47 As the maleimide functional group has two stretching modes (asymmetric at 1707 cm-1 and symmetric at 1745 cm-1) which are geometrically orthogonal,12 the relative strengths of each mode indicate its molecular orientation relative to the surface. A convenient index of the orientation of maleimide groups is the ratio of peak areas (1707 cm-1/1745 cm-1). To calculate this index for each sample, the peak areas corresponding to each of the maleimide stretching modes were deconvoluted from baseline-corrected data by fitting to multiple Gaussian peaks. The ratio of peak areas for each sample is shown in Figure 3B. As this index was not a strong function of monolayer composition, we conclude that magnitudes of (45) Atre, S. V.; Liedberg, B.; Allara, D. L. Langmuir 1995, 11, 3882. (46) Chen, S.; Li, L.; Boozer, C. L.; Jiang, S. Langmuir 2000, 16, 9287. (47) Golden, W. G. Fourier Transform Infrared Reflection-Absorption Spectroscopy. In Fourier Transform Infrared Spectroscopy: Applications to Chemical Systems; Ferraro, J. R., Basile, L. J., Eds.; Academic Press: Orlando, 1985; Vol. 4, p 315. (48) Tolstoy, V. P.; Chernyshova, I. V.; Skryshevsky, V. A. Handbook of Infrared Spectroscopy of Ultrathin Films; John Wiley & Sons: Hoboken, NJ, 2003. (49) Frey, B. L.; Corn, R. M. Anal. Chem. 1996, 68, 3187. (50) Nuzzo, R. G.; Dubois, L. H.; Allara, D. L. J. Am. Chem. Soc. 1990, 112, 558.

Orientations of Nematic Liquid Crystals on Surfaces

Figure 3. Spectroscopic characterization of maleimide-modified SAMs including (A) baseline corrected PM-IRRAS spectra, (B) calculated ratio of peak areas at 1707 and 1745 cm-1, (C) peak areas for the maleimide symmetric and asymmetric stretching modes, and (D) optical thicknesses of the maleimidemodified SAMs.

peak areas in Figure 3A can be used to infer the composition of the interface. Shown in Figure 3C is a plot of the magnitude of the absorbance peaks at 1707 and 1745 cm-1 as a function of monolayer composition. With

Langmuir, Vol. 21, No. 14, 2005 6455

increasing mole fractions of amine-terminated functionality in the SAM, we observe the areal density of immobilized maleimide group after SSMCC treatment to systematically increase. We also used ellipsometry to characterize the maleimide-functionalized surfaces. Shown in Figure 3D is the change in ellipsometric thickness of the SAM caused by SSMCC treatment, as a function of SAM composition. Again, a trend is observed of increasing the amount of immobilized SSMCC as a function of monolayer composition. The maximum optical thickness of 1.27 ( 0.08 nm obtained at a 100% EG3-N monolayer is similar to the known dimensions of the SSMCC spacer of 1.16 nm (obtained from Pierce, Inc.) and is consistent with monolayer coverage of the maleimide. From this series of studies, we conclude that the maleimide group was incorporated into the interface. We also conclude that the orientation of the maleimide group does not change significantly as a function of areal density and that the areal density of immobilized SSMCC is a function of changing monolayer composition. These results were supported by ellipsometry. Below, we provide evidence that these interfaces can be used to immobilize defined densities of peptides. Preparation and Characterization of Peptide-Modified Interfaces. An identical series of maleimide-modified surfaces, as characterized above, were prepared for subsequent reaction with the cysteine-containing peptides. Two parallel sets of experiments were performed, again using the 1, 5, 10, 25, 50, and 100 mol % EG3-N SAMs. One set was treated with a 250 µM solution of Src-tide, and the other was treated with a 250 µM solution of the phosphorylated peptide, or p-Src-tide (chemical structures in Figure 2). Our choice of a 3 h reaction time was guided by previously published results.12 We used cysteineterminated peptides, as they site-specifically react with surface-immobilized maleimide groups.6,12,15 Unreacted maleimide groups on the surface were quenched with 2-mercaptoethanol. As described below, we characterized these surfaces by PM-IRRAS and ellipsometry. Shown in Figure 4A and B are the baseline corrected PM-IRRAS spectra of Src-tide and p-Src-tide surfaces, respectively. The incorporation of peptide functionality at these interfaces is apparent, as both amide I (1655 cm-1) and amide II (1539 cm-1) bands are present in each series.12,49 To extract the contribution of each peptide from the measured absorbance spectra, the corresponding maleimide absorbance spectra were subtracted from each peptide (baseline-corrected) data set (Figure 4C and D). The difference spectra so-obtained highlight not only the increased magnitude of the amide I and amide II absorbance peaks as a function of monolayer composition, but also clearly show a loss of intensity at 1745 cm-1 (maleimide symmetric stretching mode) after peptide immobilization. We attribute this loss in intensity of the symmetric stretching mode to a breaking of molecular symmetry upon formation of the covalent adduct (refer to Scheme 1 depicting molecular structure). As overlapping peaks in the baseline-corrected spectra (Figure 4A and B) prevented the direct analysis of peak areas, each data set was fit to multiple Gaussian peaks. The amide I peak areas were corrected by subtracting from this the initial amide I intensity obtained from the maleimide-modified surfaces and plotted as a function of monolayer composition (Figure 5A). This correction assumes that the orientation of the amide bond formed after reaction with SSMCC does not change after peptide immobilization. We make two observations regarding the data presented in Figure 5A. First, it appears that full

6456

Langmuir, Vol. 21, No. 14, 2005

Clare and Abbott

Figure 4. Baseline-corrected PM-IRRAS spectra of (A) Src-tide- and (B) p-Src-tide-modified SAMs. Difference spectra highlighting the contribution of (C) Src-tide and (D) p-Src-tide.

surface coverage of each peptide is achieved at a SAM composition near 10% EG3-N. A 9-residue peptide is larger than the SSMCC linker, and it is likely that fewer peptide molecules per unit area correspond to full surface coverage. Second, we note that the magnitude of the Src-tide amide I absorption is greater than the corresponding p-Src-tide series and that this difference in magnitude is greatest when using SAMs rich in EG3-N (10-100%). We also performed ellipsometric measurements of the peptide-modified surfaces. The changes in optical thicknesses of the SAMs after both SSMCC treatment and peptide immobilization are shown in Figure 5B. Again, the amount of immobilized peptide is observed to be a function of monolayer composition, and there exists a difference in the total amounts of Src-tide and p-Srctide present at the interface. The maximum contribution of each peptide to the total optical thickness was calculated to be 0.78 nm (Src-tide) and 0.39 nm (p-Src-tide). Although these increments are smaller than what was observed for the addition of the SSMCC linker (1.27 nm), it is not an unphysical optical thickness for a 9-residue peptide. Others have reported the optical thickness of a larger, 17-residue peptide chemisorbed directly onto a gold surface to range from 1.10 to 2.82 nm.13 Next, we sought to understand the origin of the difference in the optical thicknesses of the Src-tide- and p-Src-tide-modified surfaces. In past studies of the peptide substrate of the MAP protein kinase, it has been observed that the solution conformation of a short peptide can change upon phosphorylation.51 We sought to determine if indeed the two (51) Stultz, C. M.; Levin, A. D.; Edelman, E. R. J. Biol. Chem. 2002, 277, 47653.

peptides in our system were immobilized at different densities due to orientational or conformational differences. Indices used to characterize protein structures at interfaces by infrared spectroscopic methods are (1) the ratio of amide I/amide II peak intensities (to monitor orientation),52 and the peak position of the amide I absorbance (to monitor secondary structure).53 We characterized these two peptide sequences using these two indices. First, inspection of the peak positions for the amide I absorption for Src-tide and p-Src-tide occurs at 1655 cm-1, leading to the conclusion that both peptides adopt a similar conformation and that each possesses some alpha helical character.53 Second, we calculated amide I/amide II peak area ratios for each peptide series and plotted these values as a function of the composition of the solution used to form the monolayer (Figure 5C). Figure 5C illustrates two points: (1) changes in the orientations of peptides do occur as a function of areal density of peptide at the interface, and (2) at maximum packing of peptides (50% and 100%) both Src-tide and p-Src-tide have similar orientations (ratio values). From these results, we conclude that at maximum packing densities, each peptide has a similar orientation. Therefore, these differences in maximum packing densities appear to reflect factors other than the orientations of the peptide. It is possible that longrange forces such as electrostatic interactions may be the origin of the differences in maximum packing densities.54 (52) Lavoie, H.; Desbat, B.; Vaknin, D.; Salesse, C. Biochemistry 2002, 41, 13424. (53) Taylor, S. E.; Desbat, B.; Blaudez, D.; Jacobi, S.; Chi, L. F.; Fuchs, H.; Schwartz, G. Biophys. Chem. 2000, 87, 63. (54) Castelino, K.; Kannan, B.; Majumdar, A. Langmuir 2005, 21, 1956.

Orientations of Nematic Liquid Crystals on Surfaces

Langmuir, Vol. 21, No. 14, 2005 6457

Figure 6. (A) Optical images (using polarized microscopy) of the nematic liquid crystal 5CB in contact with surfaces presenting a SAM formed from the EG3 thiol. (B) Depiction of the bulk orientation of 5CB near the interface (as confirmed using optical methods).

Figure 5. Characterization of peptide-modified SAMs. (A) Plot of amide I peak areas for each immobilized peptide as a function of monolayer composition. (B) Ellipsometric thicknesses for those samples. (C) Ratio of peak areas at 1655 and 1539 cm-1 for each peptide-modified surface.

At a neutral pH in bulk solution, the charge states of Src-tide and p-Src-tide are +2 and 0, respectively. From the above results, we conclude that a twocomponent SAM can be used to tune the areal density of both maleimide functional groups and immobilized peptides at an interface. Below, we describe the use of these methods to prepare samples for the study of liquid crystals in contact with surfaces presenting different areal densities of immobilized peptides. Part II. Orientations of Liquid Crystals in Contact with Peptide-Functionalized Surfaces. Next, we sought to characterize the orientations of the nematic liquid crystal 5CB in contact with the peptide-modified surfaces described above. Semi-transparent films of obliquely deposited gold films were prepared as described in Methods37,39 These nanostructured gold surfaces were then treated as outlined in Part I, such that each peptide was

immobilized on SAMs prepared from solutions having compositions of 1%, 5%, 10%, 25%, 50%, and 100% EG3N. These samples present a wide range in peptide immobilization density, and therefore we could evaluate the impact of peptide density on the orientational ordering of the liquid crystal. As a control, one-component SAMs comprised of only EG3 were prepared, such that these surfaces presented no peptide. To view the optical textures of liquid crystals in contact with these surfaces, optical cells were created using two identical surfaces placed faceto-face and separated by 12 µm. Cells were filled with the nematic liquid crystal 5CB and viewed using a polarized light microscope. An image of the optical appearance of the liquid crystal in contact with the 100% EG3 SAM is shown in Figure 6A. The sample appears uniformly dark when observed using polarized light microscopy (crossed polars), as the liquid crystal is uniformly aligned by the surface. The average orientations of mesogens in the sample are parallel to one polarizer, causing the extinction of transmitted light.38 The changes in interference colors observed after inserting a quarter-wave plate into the path of transmitted light was used to characterize the orientation of 5CB with respect to the anisotropic topography of the underlying gold film.36,55 From this analysis, we determine that 5CB in contact with 100% EG3 SAMs is uniformly aligned in the direction of maximum roughness of the gold film (Figure 6B). The chemical structure of this interface, including the orientation of the terminal functional groups presented by the SAM,35,36 and interactions between 5CB and the SAM33 dictate the preferred orientations of 5CB on these surfaces. We next investigated the orientations of 5CB in contact with surfaces presenting low areal densities of immobilized peptides. Samples of 5CB in contact with Src-tide- or p-Src-tide-modified SAMs formed from 1% EG3-N solutions were studied. When viewed immediately after preparation, the optical textures of 5CB in contact with these surfaces were nonuniform and possessed many line defects (Figure 7A and B, illustrated in 7C). We sought to determine if the liquid crystal in contact with surfaces having low areal densities of peptides was at an equilibrium state. These same samples were annealed in a 36 °C oven for 17 h. Once cooled, images were captured for each sample, and are shown in Figure 7D and E. After this annealing period, line defects were largely eliminated and the samples appeared uniformly dark when viewed under polarized light microscopy. Although bulk 5CB at 36 °C exists as an isotropic phase, past studies have established that the interfacial order of a thermotropic liquid crystal can persist at temperatures higher than the temperature at which the bulk phase becomes isotropic.56,57 Upon (55) Hartshorne, N. H.; Stuart, A. Crystals and the Polarizing Microscope, 4th ed.; Edward Arnold Publishers, Ltd.: London, 1970. (56) Als-Nielsen, J.; Christensen, F.; Pershan, P. S. Phys. Rev. Lett. 1982, 48, 1107.

6458

Langmuir, Vol. 21, No. 14, 2005

Figure 7. Optical images of 5CB (polarized microscopy) in contact with surfaces presenting low areal densities of peptides immediately after preparation (A, B), and at longer time intervals (D, E). Depiction of the orientation of mesogens near the interface at short time intervals (C) and longer time intervals (F).

Figure 8. Representative optical textures of 5CB in contact with surfaces presenting high areal densities of peptides (5100% EG3-N SAMs) immediately after preparation (A, B), and at longer time intervals (D, E). Depiction of the orientation of mesogens near the interface at short time intervals (C) and longer time intervals (F).

cooling, the interfacial order of the liquid crystal plays a central role in determining the orientation of the bulk liquid crystal.57 In our experiments, we attribute the influence of thermal annealing above the bulk nematicto-isotropic transition temperature to changes in the interfacial order of 5CB. We note here that if the samples were not heated to 36 °C (i.e., kept at ∼25 °C), the annealing process was slowed to an extent that we saw no measurable change in the defect densities over the course of days. In contrast to the EG3 SAMs, the orientation of the 5CB in contact with these peptidedecorated interfaces was parallel to the direction of minimum roughness of the underlying gold film (illustrated in Figure 7F). That is, upon reaching equilibrium, we observe that the preferred orientation of 5CB in contact with these surfaces is orthogonal to what is observed with the EG3 SAM system. These results suggest that the molecular-level organization, which defined the interactions between the liquid crystal and the EG3 SAM, has been perturbed by the peptide immobilization. The presence of the peptide also appears to increase the time required for this system to equilibrate and exhibit longrange ordering. The samples of 5CB in contact with either Src-tide or p-Src-tide at higher areal densities (5-100% EG3-N) immediately after preparation were also highly nonuniform (Figure 8A and B, illustrated in 8C). In contrast, these samples did not undergo a time-dependent annealing of defects upon heated to 36 °C for up to 100 h (Figure 8D (57) Clark, N. A. Phys. Rev. Lett. 1985, 55, 292.

Clare and Abbott

and E, illustrated in 8F). Recent simulations have predicted that particles adsorbed at interfaces can dramatically slow the dynamics of alignment of liquid crystals by surfaces.58 The simulations are used to establish a relationship between the number density of adsorbed particles and the time required to achieve equilibrium where the liquid crystal exhibits uniform ordering near the interface. Another conclusion drawn from the simulations is that above a critical areal density of adsorbed particles, the time required to achieve this state approaches infinity. Our experimental evidence lends support to this model, as higher areal densities of immobilized peptides lead to exceedingly long times required to achieve the equilibrium state with long-range ordering of 5CB. These results, when combined, suggest a correlation between the areal density of immobilized peptides and the dynamic reorganization of 5CB. We propose that measurement of relaxation times and defect densities offers the basis of new approaches to quantify the density of peptides at the interface. From the above results, we conclude that the introduction of peptides at interfaces can perturb the initial alignment of liquid crystals in contact with self-assembled monolayers supported on obliquely deposited gold substrates. We find that these samples, immediately after preparation, are not in an equilibrium state and that the time required to achieve equilibrium is related to the number density of immobilized peptides at the interface. As described below, when using these systems in studies of protein-peptide binding, we elected to work with surfaces having low areal densities of peptides such that the time required to reach equilibrium was practical. Part III. Using Liquid Crystals To Detect PeptideProtein Binding Events. Past work has demonstrated that protein binding events at nanostructured gold interfaces can induce the presence of defects in liquid crystals in contact with those surfaces.25,37,38,59,60 In Part II of this work, we determined that surfaces that present low areal densities of peptide permit 5CB to relax over time to a defect-free structure. Therefore, we postulated that surfaces presenting bound protein may prevent the relaxation of defects within the liquid crystal, because the specific binding of proteins to surface-immobilized peptides increases both the effective number and the size of biomolecules at the interface. To test this proposition, we performed measurements of the orientations of nematic liquid crystal 5CB in contact with peptide-laden interfaces both before and after treatment with a phospho-specific antibody. In addition, we report a study of the resistance of these peptide-modified SAMs to the nonspecific adsorption of proteins. Ellipsometric thickness measurements were recorded and used to independently confirm the presence/absence of bound protein. Guided by the results of the previous section, we synthesized surfaces having a low areal density of surfaceimmobilized peptides by using 1% EG3-N SAMs. Figure 9A depicts our experimental design (also see Materials and Methods). Each peptide surface, one presenting Srctide and the other presenting the phosphorylated sequence p-Src-tide, was treated with a 10 µg/mL solution of monoclonal anti-phosphotyrosine IgG for 1.5 h. As a control, an identical set of peptide surfaces was treated with a 10 µg/mL solution of anti-avidin IgG, an antibody having no affinity for either peptide sequence, also for 1.5 h. Prior to measuring the orientation of 5CB on these (58) Grollau, S.; Guzman, O.; Abbott, N. L.; de Pablo, J. J. J. Chem. Phys. 2005, 122, 024703. (59) Skaife, J. J.; Abbott, N. L. Langmuir 2000, 16, 3529. (60) Skaife, J. J.; Abbott, N. L. Langmuir 2001, 17, 5595.

Orientations of Nematic Liquid Crystals on Surfaces

Langmuir, Vol. 21, No. 14, 2005 6459

Figure 9. Schematic illustrations and optical images (polarized microscopy) that demonstrate the use of liquid crystals to amplify and report specific protein-peptide binding events occurring at interfaces. (A) Scheme of protein treatments of the peptide-modified interfaces, both specific binding (Anti-phosphotyrosine IgG) and control (Anti-avidin IgG). (B) Optical images of liquid crystal samples when viewed under polarized microscopy, immediately after preparation. (C) Optical textures of 5CB after annealing.

surfaces, we obtained independent confirmation of bound IgG using ellipsometry. We observed an increase in optical thickness of 2.3 ( 0.1 nm after placing the surface presenting p-Src-tide in contact with an aqueous solution of the monoclonal antiphosphotyrosine IgG. This peptide sequence contains a phosphotyrosine residue (see Figure 2), thus permitting binding between this peptide and the phospho-specific antibody. A small negative change in optical thickness (-0.1 ( 0.1 nm), within the error of measurement, was observed for the Src-tide-modified surface after treatment with the monoclonal anti-phosphotyrosine IgG. The lack of protein binding to the immobilized Src-tide confirms the selective binding between the phospho-specific antibody and the phosphorylated peptide sequence. Similar small changes (∼0.2 ( 0.1 nm) in optical thickness were observed when placing surfaces presenting these two peptides in contact with the anti-avidin IgG protein. These results are consistent with previous reports of ethylene glycol-containing SAMs resisting the nonspecific adsorption of proteins to interfaces.19,38 Optical images of the liquid crystal in contact with the surfaces described above, when viewed immediately after preparation under polarized microscopy, are shown in Figure 9B. As expected, each of these samples had nonuniform optical textures. However, after a 17 h annealing period at 36 °C, samples of 5CB in contact with the control surface presenting no bound antiphosphotyrosine IgG (confirmed by ellipsometry) relaxed to a uniform, defect-free structure (Figure 9C). In addition, samples of liquid crystal in contact with peptide surfaces exposed to solutions of anti-avidin IgG annealed to a defectfree state over the course of 17 h. This is consistent with the ellipsometric thickness measurements confirming no nonspecifically adsorbed protein at these interfaces. Interference studies using quarter-wave plate measure-

ments again confirmed that the orientation of 5CB was defined by the underlying surface topography. In contrast, after annealing, 5CB in contact with the p-Src-tide surface presenting bound monoclonal anti-phosphotyrosine IgG remained nonuniformly oriented, even after many weeks of annealing (Figure 9C). These results support our proposition that the specific binding of an antibody to a surface-immobilized peptide increases the effective number (and size) of adsorbed particles at the interface, and thus slows the dynamic reorganization of mesogens near the interface such that the time required for this system to reach equilibrium extends beyond the experimentally accessed time-scale. This label-free detection method of protein (antibody)-peptide binding events has potential application in the design of medical diagnostics8 and in the study of enzymatic activity.14 Prior work has demonstrated that it is possible to quantify the optical response of a liquid crystal in contact with surfaces presenting increasing amounts of antibody bound to an immobilized antigen.59 In that report, a correlation is drawn between the number of line defects per unit area and the quantity of bound protein. As described below, we sought to utilize the principles outlined in Parts I-III of this study to amplify and report protein binding events occurring on a spatially resolved array of immobilized peptides. Part IV. Imaging of Spatially Resolved Peptide Arrays Using Liquid Crystals. Surface-based assays are highly amenable for high-throughput screening (HTS), as many thousands of spatially resolved chemical or biomolecular species can be simultaneously tested for a given function.9 Liquid crystals can conceivably image such highly dense arrays, as previous work has demonstrated that a nematic liquid crystal can be used to resolve surfaces patterned with feature sizes having lateral dimensions of