Multifunctional Mixed SAMs That Promote Both Cell Adhesion and

May 29, 2008 - Gregory A. Hudalla and William L. Murphy. Langmuir 2010 26 (9), 6449-6456. Abstract | Full Text HTML | PDF | PDF w/ Links. Cover Image ...
1 downloads 0 Views 2MB Size
Download PDF
Langmuir 2008, 24, 6873-6880

6873

Multifunctional Mixed SAMs That Promote Both Cell Adhesion and Noncovalent DNA Immobilization Siyoung Choi‡ and William L. Murphy*,†,‡ Departments of Biomedical Engineering, Pharmacology, and Materials Science and Engineering, and Materials Science Program, UniVersity of Wisconsin, Madison, Wisconsin 53706 ReceiVed February 20, 2008. ReVised Manuscript ReceiVed April 7, 2008 The ability of DNA strands to influence cellular gene expression directly and to bind with high affinity and specificity to other biological molecules (e.g., proteins and target DNA strands) makes them a potentially attractive component of cell culture substrates. On the basis of the potential importance of immobilized DNA in cell culture and the well-defined characteristics of alkanethiol self-assembled monolayers (SAMs), the current study was designed to create multifunctional SAMs upon which cell adhesion and DNA immobilization can be independently modulated. The approach immobilizes the fibronectin-derived cell adhesion ligand Arg-Gly-Asp-Ser-Pro (RGDSP) using carbodiimide activation chemistry and immobilizes DNA strands on the same surface via cDNA-DNA interactions. The surface density of hexanethiol-terminated DNA strands on alkanethiol monolayers (30.2-69.2 pmol/cm2) was controlled using a backfill method, and specific target DNA binding on cDNA-containing SAMs was regulated by varying the soluble target DNA concentration and buffer characteristics. The fibronectin-derived cell adhesion ligand GGRGDSP was covalently linked to carboxylate groups on DNA-containing SAM substrates, and peptide density was proportional to the amount of carboxylate present during SAM preparation. C166-GFP endothelial cells attached and spread on mixed SAM substrates and cell adhesion and spreading were specifically mediated by the immobilized GGRGDSP peptide. The ability to control the characteristics of noncovalent DNA immobilization and cell adhesion on a cell culture substrate suggests that these mixed SAMs could be a useful platform for studying the interaction between cells and DNA.

Introduction Self-assembled monolayers (SAMs) of alkanethiols on gold are common model systems for biological and biotechnological applications because of their unique set of attributes.1–5 SAMs can be designed to be “bioinert”, and therefore resistant to nonspecific protein adsorption. They can also be terminated with a wide variety of functional groups, which enables variation in their surface reactivity, charge density, hydrophilicity, and resistance to protein binding.3,6–11 Finally, the molecular structure of SAMs results in a well-defined density of functional groups on the surface.1,12,13 This characteristic is particularly important * To whom correspondence should be addressed. Tel: 608-262-2224. Fax: 608-265-9239. E-mail: [email protected]. † Departments of Biomedical Engineering, Pharmacology, and Materials Science and Engineering. ‡ Materials Science Program. (1) Houseman, B. T.; Gawalt, E. S.; Mrksich, M. Langmuir 2003, 19, 1522– 1531. (2) Singhvi, R.; Kumar, A.; Lopez, G. P.; Stephanopoulos, G. N.; Wang, D. I.; Whitesides, G. M.; Ingber, D. E. Science 1994, 264(5159), 696–698. (3) Mrksich, M.; Whitesides, G. M. Annu. ReV. Biophys. Biomol. Struct. 1996, 25, 55–78. (4) Yamauchi, F.; Kato, K.; Iwata, H. Biochim. Biophys. Acta 2004, 1672(3), 138–147. (5) Wegner, G. J.; Lee, H. J.; Marriott, G.; Corn, R. M. Anal. Chem. 2003, 75, 4740–4746. (6) Boozer, C.; Chen, S.; Jiang, S. Langmuir 2006, 22(10), 4694–4698. (7) Herrwerth, S.; Eck, W.; Reinhardt, S.; Grunze, M. J. Am. Chem. Soc. 2003, 125(31), 9359–9366. (8) Li, L.; Chen, S.; Zheng, J.; Ratner, B. D.; Jiang, S. J. Phys. Chem. B 2005, 109(7), 2934–2941. (9) Harder, P.; Grunze, M.; Dahint, R.; Whitesides, G. M.; Laibinis, P. E. J. Phys. Chem. B 1998, 102, 426–436. (10) Bain, C. D.; Evall, J.; Whitesides, G. M. J. Am. Chem. Soc. 1989, 111, 7155–7164. (11) Nuzzo, R. G.; Dubois, L. H.; Allara, D. L. J. Am. Chem. Soc. 1990, 112, 558–569. (12) Schwartz, D. K. Annu. ReV. Phys. Chem. 2001, 52, 107–137. (13) Lee, C. Y.; Nguyen, P. C.; Grainger, D. W.; Gamble, L. J.; Castner, D. G. Anal. Chem. 2007, 79(12), 4390–4400.

in mechanistic studies of intermolecular interactions, such as receptor-ligand interactions. On the basis of these attributes, alkanethiol monolayers formed on gold-coated glass substrates have been used as model systems for exploring several biological phenomena, including cell adhesion,1,14 cell migration,15 and the analysis of intermolecular interactions (e.g., protein-protein and DNA-DNA interactions).5,16–18 SAMs derivatized with covalently immobilized peptides or proteins have been particularly useful in previous studies designed to study cell adhesion. Whitesides and co-workers initially demonstrated this capability by covalently linking the fibronectinderived peptide sequence Arg-Gly-Asp (RGD) to carboxylateterminated alkanethiol SAMs via carbodiimide activation chemistry.14 More recently, similar alkanethiol monolayers have been used to present a variety of cell adhesion peptides19 and proteins15,20,21 to cells, and these molecules can be spatially patterned on otherwise bioinert substrates for high-throughput analysis of cell-ligand interactions.22 These previous studies have demonstrated that alkanethiol monolayers on gold can be used as a platform to present specific biomolecules to cells in (14) Roberts, C.; Chen, S. C.; Mrksich, M.; Martichonok, V.; Ingber, D. E.; Whitesides, G. M. J. Am. Chem. Soc. 1998, 120, 6548–6555. (15) Liu, L.; Ratner, B. D.; Sage, E. H.; Jiang, S. Langmuir 2007, 23(22), 11168–11173. (16) Okahata, Y.; Kawase, M.; Niikura, K.; Ohtake, F.; Furusawa, H.; Ebara, Y. Anal. Chem. 1998, 70(7), 1288–1296. (17) Gong, P.; Lee, C. Y.; Gamble, L. J.; Castner, D. G.; Grainger, D. W. Anal. Chem. 2006, 78(10), 3326–3334. (18) Brockman, J. M.; Frutos, A. G.; Corn, R. M. J. Am. Chem. Soc. 1999, 121, 8044–8051. (19) Petrie, T. A.; Capadona, J. R.; Reyes, C. D.; Garcia, A. J. Biomaterials 2006, 27(31), 5459–5470. (20) Liu, X.; Jang, C. H.; Zheng, F.; Jurgensen, A.; Denlinger, J. D.; Dickson, K. A.; Raines, R. T.; Abbott, N. L.; Himpsel, F. J. Langmuir 2006, 22(18), 7719– 7725. (21) Murphy, W. L.; Mercurius, K. O.; Koide, S.; Mrksich, M. Langmuir 2004, 20(4), 1026–1030. (22) Mrksich, M. Chem. Soc. ReV. 2000, 29, 267–273.

10.1021/la800553p CCC: $40.75  2008 American Chemical Society Published on Web 05/29/2008

6874 Langmuir, Vol. 24, No. 13, 2008

a background that is inert to nonspecific protein adsorption. Therefore, SAMs have allowed investigators to mechanistically explore the effects of extracellular ligands on particular cell behaviors, including adhesion,14 migration,23 and differentiated function.24 The unique characteristics of DNA and RNA molecules have also led investigators to develop schemes for poly(nucleotide) immoblization on SAMs. For example, the ability of DNA strands to bind complementary sequences has been used as a mechanism to build biosensors for genomics applications25 and selfassembling substrates for de novo gene synthesis.26 The ability of DNA and RNA molecules to fold into complex protein-binding ligands, termed “aptamers”, has been used to create biosensors for protein detection.18,27 Furthermore, the ability of DNA strands to interact electrostatically with substrate-immobilized poly(cations) has been used as a mechanism to immobilize plasmid DNA for subsequent delivery to cells.28,29 Taken together, these properties of poly(nucleotides) make them a potentially attractive component of the aforementioned SAM-based model systems for cell biology. On the basis of the potential importance of immobilized poly(nucleotides) in cell culture applications and the well-defined characteristics of SAMs, the current study was designed to create multifunctional SAMs upon which cell adhesion and DNA immobilization could be independently modulated. Our approach immobilizes the fibronectin-derived cell adhesion ligand ArgGly-Asp-Ser-Pro (RGDSP) using carbodiimide activation chemistry, and the surface density of this peptide is controlled in an otherwise bioinert background by varying the ratio of carboxylateterminated and tri(ethylene glycol)-terminated molecules in the SAM layer. DNA strands are incorporated into the alkanethiol SAMs using a backfill method,30,31 and complementary target DNA strands are then immobilized on the substrate via cDNA-DNA interactions. We expect that controlling cell adhesion on SAMs that present target poly(nucleotides) may enable well-defined studies of the interaction between cells and poly(nucleotide) sequences, including antisense oligonucleotides, small interfering RNA (siRNA), plasmid DNA, and aptamer ligands. Substrate Design Rationale. Recently, Levicky and coworkers developed an approach to generating DNA-containing SAMs using thiol-terminated, single-stranded oligonucleotides (HS-ssDNA). A backfill method was used in this study to control the DNA surface coverage and achieve high hybridization efficiency in biosensor applications.30 To create these substrates, a gold substrate was first exposed to a solution of HS-ssDNA to promote adsorption and then incubated in a solution of mercaptohexanol (MCH)30,31 or 11-mercapto-1-undecanol (MCU)32 molecules, which competed with interactions between the oligonucleotide and the gold surface. The result, characterized (23) Yousaf, M. N.; Houseman, B. T.; Mrksich, M. Proc. Natl. Acad. Sci. U. S. A. 2001, 98(11), 5992–5996. (24) Keselowsky, B. G.; Collard, D. M.; Garcia, A. J. Proc. Natl. Acad. Sci. U. S. A. 2005, 102(17), 5953–5957. (25) Sassolas, A.; Leca-Bouvier, B. D.; Blum, L. J. Chem. ReV. 2008, 108(1), 109–139. (26) Tian, J.; Gong, H.; Sheng, N.; Zhou, X.; Gulari, E.; Gao, X.; Church, G. Nature 2004, 432(7020), 1050–1054. (27) Johnson, S.; Evans, D.; Laurenson, S.; Paul, D.; Davies, A. G.; Ferrigno, P. K.; Walti, C. Anal. Chem. 2008, 80, 978–983. (28) Pannier, A. K.; Anderson, B. C.; Shea, L. D. Acta Biomater. 2005, 1(5), 511–522. (29) Jewell, C. M.; Zhang, J.; Fredin, N. J.; Lynn, D. M. J. Controlled Release 2005, 106(1-2), 214–223. (30) Levicky, R.; Herne, T. M.; Tarlov, M. J.; Satija, S. K. J. Am. Chem. Soc. 1998, 120, 9787–9792. (31) Herne, T. M.; Tarlov, M. J. J. Am. Chem. Soc. 1997, 119, 8916–8920. (32) Lee, C. Y.; Gong, P.; Harbers, G. M.; Grainger, D. W.; Castner, D. G.; Gamble, L. J. Anal. Chem. 2006, 78(10), 3316–3325.

Choi and Murphy

via X-ray photoelectron spectroscopy (XPS), neutron reflectivity, ellipsometry, and surface plasmon resonance (SPR), was a SAM substrate containing HS-ssDNA molecules oriented perpendicular to the substrate in a background populated with either MCH or MCU molecules.17,30–32 In the current study, we used a similar backfill procedure to create DNA-containing SAMs with a background that promotes specific cell adhesion yet is inert to nonspecific cell interactions. The substrates (Figure 1) have three components:(i) a DNA molecule terminated with a 5′-hexanethiol moiety; (ii) an alkanethiol terminated with 1-carboxy, hexa(ethylene glycol) (EG6-carboxylate-terminated alkanethiol); and (iii) an alkanethiol terminated with tri(ethylene glycol) (EG3terminated alkanethiol). The DNA molecule is used as a complementary sequence for target DNA immobilization, the EG6-carboxylate-terminated alkanethiol is used to covalently immobilize an amine-terminated cell adhesion peptide, and the EG3-terminated alkanethiol creates a background that is inert to nonspecific protein and cell interactions. In this study, we focus on the macroscopic characterization of cells on the substrate. However, it is important to note that the substrates may not be entirely homogeneous, as shown in Figure 1. Some previous topological studies of mixed alkanethiol SAMs have demonstrated surface-phase separation within SAMs to form domains of each alkanethiol on the nanometer scale.33,34

Materials and Methods Preparation of DNA-Containing Alkanethiol Monolayers. Oligonucleotides were purchased from Operon Biotechnologies (Huntsville, AL). Gold-coated glass slides (EMF Corporation, Ithaca, NY) coated with 5 nm of Cr and 100 nm of Au (99.9% purity) were used as substrates. Prior to SAM formation, gold chips (0.8 cm × 0.8 cm) were sonicated in 100% ethanol for 1 min and rinsed with ethanol for 10 s to remove surface contaminants and then dried under N2. Sodium chloride, calcium chloride, Tris-HCl, ethylenediaminetetraacetic acid (EDTA), sodium hydroxide, and hydrochloric acid were purched from Fisher Scientific Inc. (Pittsburgh, PA). Various oligo(ethylene glycol)-containing alkanethiols [SH-(CH2)11(OCH2CH2)3-OH (Sigma-Aldrich, St. Louis, MO) and SH-(CH2)11(OCH2CH2)6-OCH2-COOH (Prochimia, Poland)] were diluted with DI water (18 MΩ cm) and mixed before the backfill procedure. Mixed alkanethiol solutions contain 0, 20, 50, or 100% EG6carboxylate-terminated alkanethiol and 100, 80, 50, or 0% EG3terminated alkanethiol, respectively. A previously described backfill procedure30,31 was used to generate DNA-containing alkanethiol monolayers on gold (Figure 1b) with varying DNA surface coverage. The process involved the following steps. A DNA sequence (5′AAAAAAAAAAGACGGCAGCGTGCAG3′ (cDNA), 2 µM) modified with a terminal hexanethiol moiety in a preparation buffer (1 M CaCl2, 10 mM Tris-HCl, 1 mM EDTA, pH 7.4) was adsorbed onto a gold substrate via a 12 h incubation. After DNA adsorption, samples were washed with DI H2O to remove loosely bound cDNA and then dried under N2. The resulting cDNAcontaining surfaces were incubated in solutions of alkanethiols (50 µM total alkanethiol in DI H2O) for 0-18 h to backfill the surface with alkanethiol molecules and to competitively remove cDNA that was nonspecifically adsorbed to the surface. The resulting substrates were washed with DI H2O and dried under N2. Quantifying Surface Coverage of cDNA on cDNA-Containing Alkanethiol Monolayers. The density of cDNA on SAM surfaces was characterized by removing cDNA from the surface via competitive displacement with 2-mercaptoethanol (Fisher Scientific Inc., Pittsburgh, PA). cDNA-containing substrates were immersed in 12 mM 2-mercaptoethanol in TE buffer (10 mM Tris-HCl, 1 mM EDTA, pH 7.4) to displace cDNA from the surface. After 24 h of (33) Tamada, K.; Hara, M.; Sasabe, H.; Knoll, W. Langmuir 1997, 13, 1558– 1566. (34) Stranick, S. J.; Parikh, A. N.; Tao, Y.-T.; Allara, D. L.; Weiss, P. S. J. Phys. chem. 1994, 98, 7636–7646.

Multifunctional Mixed SAMs

Langmuir, Vol. 24, No. 13, 2008 6875

Figure 1. Formation of cDNA-containing alkanethiol monolayers. (a) A DNA sequence terminated with a hexanethiol linker is adsorbed onto a gold surface overnight in preparation buffer (1 M CaCl2, 10 mM Tris-HCl, 1 mM EDTA, pH 7.4). (b) A mixture of an EG3-terminated alkanethiol and an EG6-carboxylate-terminated alkanethiol (50 µM total alkanethiol in DI H2O) is added to backfill the SAM layer and competitively replace cDNA on the gold substrate. (c) An amine-terminated GGRGDSP peptide is conjugated to the carboxylate groups on the surface via a carbodiimide activation reaction. (d) Target DNA is noncovalently bound to the surface via a cDNA interaction.

incubation, the solution containing displaced cDNA was analyzed using a Quant-iT OliGreen ssDNA assay kit (Invitrogen, Carlsbad, CA). The fluorescence measured at 520 nm was converted to the amount of cDNA using standard curves prepared with known concentrations of cDNA. This method was chosen on the basis of a previous study in which Dermers et al. showed that mercaptoethanol competitively displaced thiol-terminated oligonucleotides from gold surfaces within 10 h of exposure.35 Characterization of Noncovalent DNA Immobilization. cDNAcontaining alkanethiol SAMs were then incubated with target DNA to allow for noncovalent immobilization. SPR was used to measure binding of target DNA (5′GAGCTGCACGCTGCCGTC3′) to cDNAcontaining, EG3-terminated alkanethiol monolayers. SAM layers were formed on gold substrates (SIA AU sensor, Biacore, Sweden) by exposing the substrates to 2 µM cDNA in preparation buffer overnight. After the adsorption of cDNA, samples were washed with DI H2O and dried under N2. For backfill, substrates were then immersed in 50 µM EG3-terminated alkanethiol in DI H2O. After 1 h of incubation, substrates were removed from the solution, washed thoroughly in DI H2O, and dried under N2. A Biacore 2000 spectrometer (Biacore, Sweden) was used to investigate DNA-DNA binding. All experiments were performed at 25 °C, and the flow rate was 30 µL/min unless otherwise noted. To control target DNA binding, different concentrations of target DNA were prepared in (35) Demers, L. M.; Mirkin, C. A.; Mucic, R. C.; Reynolds, R. A., 3rd; Letsinger, R. L.; Elghanian, R.; Viswanadham, G. Anal. Chem. 2000, 72(22), 5535–5541.

TE buffer containing 1 M CaCl2 or 1 M NaCl. All experiments were repeated in triplicate on different substrates. The association rate constant (ka) and dissociation rate constant (kd) values were calculated after fitting the sensorgram data using a simple model for 1:1 (Langmuir binding) with BIAevaluation software (version 4). DNA-DNA association constant (KA) values were evaluated by dividing ka by kd. Binding experiments were performed at three separate flow rates (5, 15, and 75 µL/min) to assess the potential impact of mass transfer limitations on the binding constants. To confirm target DNA binding, 10 µL of target DNA (10 µM or 100 µM in preparation buffer) was “spotted” onto EG3-terminated alkanethiol monolayers and cDNA-containing, EG3-terminated alkanethiol monolayers. After 1 h of incubation at 37 °C, the substrates were washed thoroughly in DI H2O and dried under N2. The substrates were then incubated with the Quant-iT PicoGreen dsDNA assay reagent (Invitrogen, Carlsbad, CA) for 30 min and imaged using an Olympus IX51 inverted fluorescence microscope (Olympus, Center Valley, PA) at 1.25× magnification. Preparation of SAMs Presenting Both DNA and the GGRGDSP Peptide. To prepare SAMs presenting both a cDNA sequence and a cell adhesion peptide (Figure 1c), we performed carbodiimide activation chemistry on a mixed-SAM surface. The fibronectinderived cell adhesion peptide GGRGDSP was synthesized with standard solid-phase peptide synthesis using Fmoc chemistry. Amino acids and the Fmoc-Rink amide MBHA resin were purchased from Nova Biochem (La Jolla, CA). Upon resin cleavage, the peptide was

6876 Langmuir, Vol. 24, No. 13, 2008 desalted via dialysis, lyophilized, and characterized via matrixassisted laser desorption/ionization-time-of-flight mass spectrometry (MALDI-TOF MS) (Bruker, Billerica, MA) for future use. In peptide immobilization experiments, the preparation of SAM substrates was modified to include a carboxylate-terminated alkanethiol. Specifically, during the backfill procedure EG3-terminated and EG6-carboxylate-terminated alkanethiols were mixed in DI H2O at different ratios to regulate the amount of carboxylate on the resulting SAMs. The total alkanethiol concentration was held constant at 50 µM. After 5 h of incubation, samples were washed with DI H2O and dried under N2. To immobilize the GGRGDSP peptide, carboxylatecontaining SAMs were first exposed to an aqueous solution with 250 mM of 1-ethyl-3-[3-(dimethylamino)propyl]carbodiimide (EDC) (Pierce, Rockford, IL) and 100 mM of N-hydroxysuccinimide (NHS) (Pierce, Rockford, IL) at room temperature. After 25 min, the samples were washed with DI H2O and dried with N2. These SAMs, which contained activated NHS esters, were then reacted for 1 h with 2 mM GGRGDSP peptide in PBS (pH 7.4) at room temperature. The resulting samples presenting both cDNA and GGRGDSP peptide were hybridized with target DNA and used for cell adhesion experiments. Each reaction step was confirmed by polarization modulation infrared reflection-absorption spectroscopy (PM-IRRAS) using a Nicolet Magna-IR 860 FTIR spectrometer with a photoelastic modulator (Hinds Instruments, Hillsboro, OR) and a synchronous sampling demodulator (GWC Technologies, Madison, WI). All spectra were recorded in the range of 1000-3000 cm-1. The modulation was centered at 1500 and 2500 cm-1 at an incident angle of 83°. To confirm GGRGDSP peptide conjugation, the amide I peak area was calculated by Simpson’s rule, which is a method for approximating definite integrals using quadratic polynomials. Integral points were selected after subtracting spectra obtained before GGRGDSP peptide conjugation from spectra obtained after conjugation. To analyze nonspecific interactions between target DNA and GGRGDSP peptide on the SAM surface, we performed an additional control experiment in which SPR was used to evaluate target DNA binding to a substrate, which was prepared by backfill with a mixed alkanethiol solution containing 10% EG6-carboxylateterminated alkanethiol and 90% EG3-terminated alkanethiol. Cell Adhesion on SAMs Presenting Target DNA and the GGRGDSP Peptide. Endothelial cells were next seeded onto various alkanethiol SAMs to characterize cell adhesion on substrates presenting both an immobilized target DNA sequence and a cell adhesion peptide. C166-GFP, a mouse endothelial cell line, was obtained from American Type Culture Collection (ATCC, Rockville, MD) and grown in Dulbecco’s Modified Eagle’s Medium (DMEM) supplemented with 10% calf serum and 1% penicillin/streptomycin. All cultures were grown in 5% CO2 at 37 °C in a humidified incubator. Cells were seeded onto SAMs at a density of 20 000 cells/cm2 in serum-free medium. After 5 h of incubation at 37 °C, the substrates were removed and gently washed with serum-free medium. Cells were then removed from the substrates via treatment with 0.05% trypsin/EDTA, and the removed cells were characterized using the CyQUANT cell proliferation assay kit (Invitrogen, Carlsbad, CA). Fluorescence measured at 520 nm was converted to cell number using standard curves, relating fluorescence intensity to cell number. To verify that cell adhesion was mediated by the GGRGDSP peptide, we performed a competition experiment using a soluble GGRGDSP peptide. Then, 20 000 cells in 1 mL of serum-free medium were incubated in suspension for 15 min at 37 °C with different concentrations of soluble GGRGDSP peptide. The cell suspension and soluble GGRGDSP peptide were seeded onto GGRGDSP peptide-containing SAMs, and SAMs were prepared by backfilling with a mixture of 10% EG6-carboxylate-terminated alkanethiol and 90% EG3-terminated alkanethiol solution (50 µM total alkanethiol in DI H2O), followed by reaction with GGRGDSP peptide as described above. After 3 h of incubation at 37 °C, the substrates were gently washed with serum-free medium and imaged using an inverted fluorescence microscope at 4× and 10× magnification. After imaging, cells were removed from the substrates via treatment with 0.05% trypsin/EDTA, and the CyQUANT assay described above was used to quantify the number of cells.

Choi and Murphy

Figure 2. Surface densities of cDNA as a function of backfill time in mixed alkanethiol solutions. Mixed alkanethiol solutions contain 0, 20, 50, or 100% EG6-carboxylate-terminated alkanethiol and 100, 80, 50, or 0% EG3-terminated alkanethiol, respectively. Values represent the mean ( standard deviation (n ) 3).

Results and Discussion Surface Coverage of cDNA on Alkanethiol Monolayers. The cDNA density on SAM surfaces can be varied using a modification of the backfill approach described previously by Herne et al.31 Surface densities of cDNA (HS-(CH2)65′AAAAAAAAAAGACGGCAGCGTGCAG3′) on mixed alkanethiol monolayers were quantified using a mercaptoethanol replacement method. cDNA immobilization exhibits a maximum density with no alkanethiol backfill (69.2 ( 3.3 pmol/cm2). This value is consistent with the previous report from Steel et al. in which the single-stranded oligonucleotide density on gold substrates prior to backfill was 49.9-66.4 pmol/cm2.36 When cDNA-containing gold substrates were incubated with a 100% EG3-terminated alkanethiol, the surface cDNA density decreased with increasing incubation time (Figure 2). This change in cDNA density can be attributed to the competitive displacement of cDNA molecules by EG3-terminated alkanethiol molecules, as described previously.31 After 18 h of backfill time, the surface density of cDNA decreased by approximately 2-fold (30.2 ( 3.3 pmol/ cm2) when compared to the initial cDNA density (69.2 ( 3.3 pmol/cm2). These results show similar trends when compared to a previous XPS study performed on DNA-containing alkanethiol monolayers by Lee et al.37 In this previous study, DNA monolayers on gold were exposed to 50 µM SH-(CH2)11(OCH2CH2)4-OH solution for various exposure times, and DNA was gradually displaced from the gold surface. Therefore, our results corroborate previous reports that demonstrate the competitive replacement of DNA from the surface by more strongly bound alkanethiol molecules. They also indicate that the resulting substrates represent an adaptable platform for noncovalent DNA immobilization on alkanethiol SAMs. cDNA-containing gold substrates backfilled with a mixture of an EG3-terminated alkanethiol and an EG6-carboxylate-terminated alkanethiol showed a decreased displacement of cDNA molecules when compared to backfill with 100% EG3-terminated alkanethiol (Figure 2). The extent of displacement was dependent on the amount of carboxylate included in the alkanethiol solution. Previous studies by Bain et al. report that the surface composition of mixed alkanethiols on SAMs are not identical to the corresponding alkanethiol composition in solution.38,39 Therefore, it is likely that the EG6-carboxylate-terminated alkanethiol and the EG3-terminated alkanethiol used in the current study interact with the surface differently. In addition, He et al. demonstrated (36) Steel, A. B.; Levicky, R. L.; Herne, T. M.; Tarlov, M. J. Biophys. J. 2000, 79(2), 975–981. (37) Lee, C.-Y.; Gamble, L. J.; Grainger, D. W.; Castner, D. G. Biointerphases 2006, 1(2), 82–92. (38) Bain, C. D.; Evall, J.; Whitesides, G. M. J. Am. Chem. Soc. 1989, 111, 7155–7164. (39) Bain, C. D.; Whitesides, G. M. J. Am. Chem. Soc. 1988, 110, 6560–6561.

Multifunctional Mixed SAMs

that carboxylic acid-terminated SAMs are fully ionized above pH 7 using a chemical force titration method,40 and all experiments described here were performed at pH 7.4. Therefore, electrostatic repulsion between negatively charged carboxylate groups and cDNA adsorbed on gold substrates may have a negative impact on the adsorption of EG6-carboxylate-terminated alkanethiol in the current study. Target DNA Binding on cDNA-Containing Alkanethiol Monolayers. Target DNA binds specifically to cDNA-containing SAMs, and binding can be modulated by varying the target DNA concentration and solution characteristics. SPR was used to detect target DNA binding under multiple buffer conditions on cDNAcontaining EG3-terminated alkanethiol monolayers. All samples for SPR measurement were backfilled with 100% EG3-terminated alkanethiol for 1 h, resulting in a surface cDNA density of 41.3 ( 4.6 pmol/cm2 (Figure 2). Results indicate that binding is dependent on the target DNA concentration in solution and the buffer characteristics (Figure 3). The amount of target DNA bound increases with increased target DNA concentration in solution (Figure 3a,b), as expected. Therefore, the variation of the concentration of target DNA provides a mechanism to control immobilized DNA density. This result is in general agreement with a previous report by Okahata et al. in which a quartz crystal microbalance (QCM) was used to demonstrate that target DNA binding increased with increased target DNA concentration in 0.2 M NaCl solution.16 In a 1 µM target DNA solution, the DNA-DNA association constant (KA) values were (333 ( 27.1) × 106 M-1 with 1 M CaCl2 and (32.1 ( 4.8) × 106 M-1 with 1 M NaCl, respectively. These values, obtained for a 15mer cDNA interaction, are lower than Okahata’s aforementioned results (570 × 106 M-1) with a 20mer cDNA interaction, which indicates that cDNA binding becomes stronger with increasing sequence length of DNA, as expected. Another factor that may explain this difference in KA is the DNA density on the substrate. Peterson et al. found that the efficiency of cDNA binding decreased with increased cDNA density on a substrate, which is likely due to increased electrostatic repulsive interactions and steric hindrance.41 Our cDNA density on the substrate (41.3 ( 4.6 pmol/cm2) was nearly 2-fold higher than that of Okahata’s result (20.4 pmol/cm2), which may partially explain our lower KA value. KA values were calculated only for experiments using 1 µM target DNA concentrations. This is due to the influence of mass transfer on rate constants at lower target DNA concentrations, as demonstrated in the Supporting Information (Supplementary Figure 1). It is noteworthy that when the concentration of target DNA is held constant the overall binding of target DNA is higher in buffer containing 1 M CaCl2 when compared with that in buffer containing 1 M NaCl (Figure 3a,b). At a target DNA concentration of 1 µM, the KA in buffer with 1 M CaCl2 ((333 ( 27.1) × 106 M-1) was nearly 10-fold higher when compared to the KA in buffer with 1 M NaCl ((32.1 ( 4.8) × 106 M-1). This effect may be explained by a more effective reduction of repulsive interactions between cDNA and target DNA in the presence of divalent cations. Theoretical models of interactions between salt ions and DNA indicate that divalent cations reduce the DNA charge more than do monovalent cations.42–44 Therefore, the presence of divalent cations may enhance cDNA binding by reducing repulsive interactions between cDNA and target DNA. (40) He, H.; Huang, W.; Zhang, H.; Li, Q. G.; Li, S. F. Y.; Liu, Z. F. Langmuir 2000, 16, 517–521. (41) Peterson, A. W.; Heaton, R. J.; Georgiadis, R. M. Nucleic Acids Res. 2001, 29(24), 5163–5168. (42) Manning, G. S. Macromolecules 2001, 34, 4650–4655. (43) Manning, G. S. Q. ReV. Biophys. 1978, 11, 179–246. (44) Patra, C. N.; Yethiraj, A. Biophys. J. 2000, 78(2), 699–706.

Langmuir, Vol. 24, No. 13, 2008 6877

Figure 3. SPR analysis of target DNA hybridization on cDNA-containing, EG3-terminated alkanethiol monolayers. (a) Target DNA in TE buffer + 1 M CaCl2 with a concentration of 1000, 100, or 10 nM was allowed to flow over a cDNA-containing, EG3-terminated alkanethiol monolayer for 10 min, followed by the injection of buffer only. (b) Target DNA in TE buffer + 1 M NaCl with a concentration of 1000, 500, 100, or 50 nM was allowed to flow over a cDNA-containing, EG3-terminated alkanethiol monolayer for 10 min, followed by the injection of buffer only (flow rate ) 30 µL/min). (c) SPR analysis of target DNA binding and subsequent removal from cDNA-containing, EG3-terminated alkanethiol monolayers. (i) 1 µM nontarget DNA (5′GGATCTTCACCTAGATCCT3′) was allowed to flow over the surface to test nonspecific binding. (ii) 1 µM of target DNA was allowed to flow over the surface, demonstrating specific binding to the surface. (iii) 1 µM of soluble cDNA was allowed to flow over the surface, resulting in the competitive removal of bound target DNA (TE buffer + 1 M CaCl2, flow rate ) 10 µL/min). (d) Fluorescence analysis of target DNA spotted onto SAM substrates. Shown are an EG3-terminated alkanethiol monolayer exposed to 100 µM target DNA (left) and a cDNA-containing, EG3-terminated alkanethiol monolayer exposed to 10 µM target DNA (right). Scale bar ) 1 mm.

The DNA immobilization efficiency was calculated by dividing the density of bound target DNA (characterized via SPR) by the density of cDNA in the SAM (characterized via the aforementioned 2-mercaptoethanol replacement method). More specifically, the bound target DNA density was estimated by using the resonance unit (RU) value measured after 10 min of target DNA binding, the known fluid channel area (1.2 mm2), and the relation 1 RU value ) 1 pg/mm2 determined in a previous study.45,46 (45) Cho, Y. K.; Kim, S.; Kim, Y. A.; Lim, H. K.; Lee, K.; Yoon, D.; Lim, G.; Pak, Y. E.; Ha, T. H.; Kim, K. J. Colloid Interface Sci. 2004, 278(1), 44–52.

6878 Langmuir, Vol. 24, No. 13, 2008

Results indicate that the target DNA binding efficiency in buffer with 1 M CaCl2 at a concentration of 1 µM (40.7%) was more than 2-fold higher than the binding efficiency in buffer with 1 M NaCl at a concentration of 1 µM (18%). The binding efficiency measured in buffer with 1 M NaCl was in good agreement with previous results from Lee et al., which measured a 20% binding efficiency for target DNA binding on DNA-containing alkanethiol monolayers backfilled with EG4-terminated alkanethiol for 1 h in buffer with 1 M NaCl.37 The increased binding efficiency in 1 M CaCl2 buffer suggests that it may be possible to modulate noncovalent DNA immobilization by changing the identity of salts in solution, even in the case of equivalent surface cDNA density and target DNA concentration. To confirm that target DNA binding to the surface-bound cDNA was specific, we next compared the binding of complementary target DNA (5′GAGCTGCACGCTGCCGTC3′) and noncomplementary target DNA (5′GGATCTTCACCTAGATCCT3′). The binding of noncomplementary target DNA to cDNAcontaining, EG3-terminated alkanethiol monolayers in buffer with 1 M CaCl2 was not detected (Figure 3c(i)). This result indicates that the EG3-terminated alkanethiol SAM background effectively prevents nonspecific DNA binding on the surface. After the injection of 1 µM target DNA in 1 M CaCl2 buffer, target DNA binding onto cDNA-containing, EG3-terminated alkanethiol monolayers reached 1103 RU, which corresponded to a density of 16.8 pmol/cm2 (Figure 3c(ii)). To confirm the specificity of target DNA binding, 1 µM soluble cDNA (5′GACGGCAGCGTGCAGCTC3′) was then injected and allowed to flow for 10 min, and the RU value was significantly reduced (Figure 3c(iii)). This result demonstrates that soluble cDNA can competitively remove bound target DNA. Target DNA binding was also demonstrated by spotting a solution of target DNA onto SAM substrates (Figure 3d). Results demonstrate that an EG3-terminated alkanethiol does not bind to target DNA. In contrast, target DNA (10 µM) binds to cDNAcontaining, EG3-terminated alkanethiols only in the “spotted” region. Taken together, these results reveal that target DNA is noncovalently bound to the surface via a cDNA interaction and that target DNA immobilization can be localized to specific regions on the SAM substrate. GGRGDSP Ligand Presentation on cDNA-Containing Alkanethiol SAMs. The GGRGDSP peptide could be conjugated to carboxylate-containing SAMs via carbodiimide activation chemistry, and the amount of immobilized peptide was dependent on the amount of carboxylate present on the surface. PM-IRRAS spectra of the EG6-carboxylate-terminated alkanethiol monolayer (Figure 4a(i)) showed three dominant peaks that can be attributed to the C-O-C stretch from the EG units of the alkanethiols (1128 cm-1) and symmetric and asymmetric stretching of the alkyl CH2 of the alkanethiols (2867 and 2919 cm-1, respectively). The peak at 2919 cm-1 indicates well-ordered packing of the alkyl chains with a low density of gauche defects.9 PM-IRRAS spectra of cDNA monolayers prior to backfill (Figure 4a(ii)) included bands at 1108 and 1243 cm-1, which represent the symmetric and asymmetric PO2- stretches of the DNA phosphodiester backbone. In addition, peaks associated with CdO, CdN, and exocyclic NH2 bending in the DNA bases appeared in the region from 1600 to 1750 cm-1.47 After the backfill of a cDNA-containing substrate with an EG6-carboxylate-terminated alkanethiol solution, peaks from both EG6-carboxylate-terminated alkanethiol and cDNA monolayers were apparent, demonstrating (46) Stenberg, E.; Persson, B.; Roos, H.; Urbaniczky, C. J. Colloid Interface Sci. 1991, 143, 513–526. (47) Brewer, S. H.; Anthireya, S. J.; Lappi, S. E.; Drapcho, D. L.; Franzen, S. Langmuir 2002, 18, 4460–4464.

Choi and Murphy

Figure 4. (a) PM-IRRAS spectra of cDNA-containing, EG6-carboxylateterminated alkanethiol monolayer formation. (i) EG6-carboxylateterminated alkanethiol monolayers showing peaks associated with alkanethiols (9), the C-O-C stretch from the EG6 and CH2 from the alkyl chain. (ii) cDNA monolayer prior to backfill showing peaks associated with DNA bases (b), PO2- from the DNA phosphodiester backbone, and CdO and CdN stretching, and exocyclic NH2 bending. (iii) cDNA-containing alkanethiol monolayer backfilled with 100% EG6carboxylate-terminated alkanethiol showing peaks associated with both DNA (b) and the alkanethiol (9). (iv) The monolayer in iii after reaction with GGRGDSP peptide, showing the peak associated with the amide I band from the peptide (2). (b) Integrated areas of amide I peaks after GGRGDSP peptide immobilization on SAMs prepared with different amounts of EG6-carboxylate-terminated alkanethiol. Values represent the mean ( standard deviation (n ) 3).

the formation of mixed SAMs (Figure 4a(iii)). A major difference between the spectra of adsorbed cDNA (Figure 4a(ii)) and cDNAcontaining, EG6-carboxylate-terminated alkanethiol is the band in the region from 2800 to 3000 cm-1. The asymmetric alkyl CH2 stretching mode at 2926 cm-1 suggests less-ordered packing of the alkyl chains and more gauche defects in the cDNAcontaining SAM when compared with the EG6-carboxylateterminated alkanethiol monolayer. The symmetric alkyl CH2 stretching mode at 2865 cm-1 also indicates that some of the alkyl chains of the EG6-carboxylate-terminated alkanethiol have gauche conformations.9 This result indicates that cDNAcontaining alkanethiol monolayers backfilled with EG6-carboxylate-terminated alkanethiol have slightly less ordered packing than does an EG6-carboxylate-terminated alkanethiol monolayer without cDNA. The covalent immobilization of the GGRGDSP peptide to carboxylate-containing SAMs results in the appearance of an amide I peak at 1674 cm-1 (Figure 4a, iv).48 To control the amount of immobilized peptide, the GGRGDSP peptide was conjugated to substrates that were prepared by backfill with mixed alkanethiol solutions containing 0, 20, 50, or 100% EG6carboxylate-terminated alkanethiol and 100, 80, 50, or 0% EG3terminated alkanethiol, respectively. The amide I peak area after the immobilization of GGRGDSP peptide increased with increasing carboxylate density on cDNA-containing alkanethiol monolayers (Figure 4b). Taken together, these results indicate that the GGRGDSP peptide density can be controlled by modulating the carboxylate density on cDNA-containing SAMs. The direct measurement of the conversion efficiency of the NHS ester to amide is difficult using PM-IRRAS because the NHS ester peaks overlap with the peaks from DNA bases in the range (48) Frey, B. L.; Corn, R. M. Anal. Chem. 1996, 68, 3187–3193.

Multifunctional Mixed SAMs

Figure 5. (a) Adhesion of C166-GFP endothelial cells to various alkanethiol monolayers in a serum-free medium. Mixed alkanethiol solutions contain 0, 0.5, 1, 5, or 20% EG6-carboxylate-terminated alkanethiol and 100, 99.5, 99, 95, or 80% EG3-terminated alkanethiol. Each surface containing EG6-carboxylate-terminated alkanethiols was reacted with the GGRGDSP peptide prior to cell seeding, as described in the text. Values represent the mean ( standard deviation (n ) 3). * indicates statistical significance relative to the condition with 0% EG6carboxylate-terminated alkanethiol during backfill (p < 0.05). (b) Images of C166-GFP endothelial cells on various alkanethiol monolayers. All images were taken 5 h after cell seeding. Cell adhesion on monolayers presenting 100% EG3-terminated alkanethiol (i), 100% cDNA (ii), and cDNA backfill with 0% (iii), 0.5% (iv), 1% (v), 5% (vi), and 20% (vii) EG6-carboxylate-terminated alkanethiol. (left images, 4× magnification; right images, 10× magnification; scale bar ) 100 µm).

Langmuir, Vol. 24, No. 13, 2008 6879

from 1600 to 1750 cm-1. We speculate that the coupling efficiency of the GGRGDSP peptide may be similar to that reported in a previous study in which poly(L-lysine) was immobilized on carboxylate-terminated SAMs using carbodiimide activation chemistry.48 Importantly, the immobilized GGRGDSP peptide had a minimal influence on target DNA binding, as evaluated by SPR (Supplementary Figure 2). More specifically, there is minimal nonspecific binding of nontarget DNA to GGRGDSP peptide-containing SAMs, and the minimal nonspecific binding is much lower when compared with the binding of target DNA to the same surface (Supplementary Figure 2). A significant portion of the amide I peak area can be attributed to amide bonds within GGRGDSP peptide and the covalent linkage between the EG6-carboxylate-terminated alkanethiol substrate and the peptide. Another conceivable contributor to the amide I peak area may be the formation of amide bonds between activated esters on the surface and DNA bases. For example, Huang et al. immobilized DNA on ω-mercaptoundecanoic acid (MUA) SAMs using carbodiimide activation chemistry.49 However, the target DNA concentration used in our current study (6 µg/mL) is below the concentration shown by Huang et al. to result in DNA immobilization (>10 µg/mL). In addition, there is indirect evidence in this article that suggests that this type of “side reaction” is not a significant contributor. First, if there were significant interactions between cDNA in the SAM and activated esters, then the cDNA interaction with target DNA would likely be strongly affected. Our observation of target DNA immobilization on cDNA-containing substrates suggests that the cDNA on the surface is free to interact with target DNA. Second, target DNA binding after the immobilization of the GGRGDSP peptide was not significantly different when compared to the target DNA binding to surfaces without conjugated peptide (Supplementary Figure 2). This result indicates that if there is a side reaction involving target DNA conjugation with NHS esters on the surface it is minimal and not likely to influence our results. C166-GFP Cell Binding on DNA-Containing SAMs Presenting the GGRGDSP Ligand. The GGRGDSP ligand promotes C166-GFP endothelial cell adhesion when presented on a DNA-containing SAM surface, and the adherent cell density is dependent on the ligand density on the surface (Figure 5). Cells showed low levels of cell adhesion on EG3-terminated alkanethiol SAMs (1260 ( 286 cells/cm2). cDNA-containing SAMs prepared without backfill promoted low levels of cell adhesion (1141 ( 901 cells/cm2). In contrast, SAMs prepared with immobilized target DNA and containing the covalently linked GGRGDSP peptide showed high levels of cell adhesion, and the adhesion density increased with increased GGRGDSP ligand density. In these experiments, the GGRGDSP peptide was conjugated to substrates, which were prepared by backfill with mixed alkanethiol solutions containing 0, 0.5, 1, 5, or 20% EG6carboxylate-terminated alkanethiol and 100, 99.5, 99, 95, or 80% EG3-terminated alkanethiol. The surfaces were then reacted with the GGRGDSP peptide and seeded with C166-GFP endothelial cells. The cell density significantly increases as the amount of EG6-carboxylate-terminated alkanethiol present in solution during backfill increases (Figure 5a), and the cell density was more than 9-fold higher on SAMs prepared via backfill with 5% EG6carboxylate-terminated alkanethiol (15313 ( 3922 cells/cm2) when compared to SAMs prepared via backfill with 100% EG3terminated alkanethiol (1672 ( 699 cells/cm2). In addition, cells on GGRGDSP peptide-containing SAMs were less rounded when compared to cells on monolayers that did not present the (49) Huang, E.; Zhou, F.; Deng, L. Langmuir 2000, 16(7), 3272–3280.

6880 Langmuir, Vol. 24, No. 13, 2008

Choi and Murphy

GGRGDSP peptide (Figure 5b). This result is in general agreement with a previous report by Roberts et al. in which the RGD peptide density was varied to demonstrate that cell attachment and spreading were affected by RGD peptide density on mixed alkanethiol monolayers.14 They reported that cell attachment was increased with increasing RGD peptide density, followed by a plateau in cell attachment at relatively low peptide density. Similarly, our results indicate that cell adhesion increases with increasing GGRGDSP ligand density before reaching a saturating density (Figure 5a). These results demonstrate that the cell adhesion density on DNA-containing SAMs can be modulated by introducing variable amounts of a fibronectin-derived cell adhesion peptide. It is also noteworthy that cell spreading is influenced by RGDSP density on the substrate, as demonstrated qualitatively (Figure 5b). To determine whether cell binding to DNA-containing SAMs was specifically mediated by the GGRGDSP peptide, we next demonstrated that the soluble GGRGDSP peptide could compete with cell adhesion to DNA-containing SAMs containing immobilized GGRGDSP peptide. As the concentration of soluble GGRGDSP peptide increased, there was a corresponding decrease in the number of cells adhered to DNA-containing SAMs presenting immobilized GGRGDSP peptide (Figure 6a). When C166-GFP cells were seeded in a solution containing 20 µM GGRGDSP peptide, the density of adherent cells was 83 ( 36% of the density observed in the absence of a soluble ligand, and the cell morphology changed from a well-spread phenotype to a more rounded phenotype (Figure 6b). When cells were seeded in a solution containing 2 mM soluble GGRGDSP peptide, the cell density decreased to 35.5 ( 8.0% of the cell density observed in the absence of soluble GGRGDSP peptide (Figure 6a). In addition, cells on these substrates did not demonstrate spreading (Figure 6b). Taken together, these results indicate that cell adhesion to DNA-containing SAMs is mediated by the GGRGDSP peptide.

Conclusions SAM substrates that present immobilized poly(nucleotides) in a bioinert background have the potential for use as controlled cell culture substrates, biosensors, and DNA delivery platforms. Here we have developed multifunctional mixed alkanethiol SAM substrates that can promote both peptide-mediated cell adhesion and DNA immobilization via cDNA interactions. The surface coverage of cDNA on alkanethiol monolayers was controlled using a backfill method, and target DNA binding on cDNAcontaining SAMs was regulated by varying the concentration of target DNA and the solution characteristics during binding. Immobilized target DNA could be removed from the surface by adding soluble cDNA, indicating that target DNA immobilization was mediated by noncovalent cDNA interactions. The fibronectinderived cell adhesion ligand GGRGDSP was covalently linked to carboxylate groups on DNA-containing SAM substrates using carbodiimide activation chemistry, and the peptide density was proportional to the amount of carboxylate present during the preparation of the SAM surface. C166-GFP endothelial cells adhered to mixed SAM substrates, and cell adhesion and spreading were modulated by varying the immobilized GGRGDSP peptide density on mixed SAM substrates. The ability to control the characteristics of noncovalent DNA immobilization and cell adhesion on an otherwise bioinert substrate suggests that these mixed SAMs could be a useful platform for studying the interaction between cells and DNA.

Figure 6. C166-GFP endothelial cell adhesion on GGRGDSP peptidecontaining SAMs. (a) Competition with cell adhesion via soluble [GGRGDSP] peptide indicates that C166-GFP endothelial cells adhere to SAMs via interaction with the immobilized GGRGDSP peptide. Values represent the mean ( standard deviation (n ) 6). * indicates statistical significance relative to the condition with 0 µM soluble peptide (p < 0.05). (b) Images of C166-GFP endothelial cell on SAMs. Cells were seeded after treatment with variable concentrations of soluble [GGRGDSP] peptide. Images demonstrate the competition with cell adhesion and spreading with increasing concentration of soluble [GGRGDSP] peptide, indicating that cells adhere to these substrates via substrate-immobilized GGRGDSP peptide. All images were taken 3 h after cell seeding. Cell adhesion on monolayers presenting 0 µM (i), 20 µM (ii), 200 µM (iii), and 2000 µM (iv) soluble [GGRGDSP] peptide. (left images, 4× magnification; right images, 10× magnification; scale bar ) 100 µm).

Supporting Information Available: Vibrational frequency assignments and SPR analyses. This material is available free of charge via the Internet at http://pubs.acs.org. Acknowledgment. We acknowledge financial support from the National Institutes of Health (R21HL084547). LA800553P