11348
Langmuir 2004, 20, 11348-11354
Micropatterning of DNA-Tagged Vesicles Brigitte Sta¨dler,† Didier Falconnet,† Indriati Pfeiffer,‡ Fredrik Ho¨o¨k,‡ and Janos Vo¨ro¨s*,† BioInterfaceGroup, Laboratory for Surface Science and Technology, Department of Materials, Swiss Federal Institute of Technology (ETH) Zurich, CH-8093 Zurich, Switzerland, and Department of Applied Physics, Chalmers University of Technology and Go¨ teborg University, SE-41296 Gothenburg, Sweden Received July 14, 2004. In Final Form: September 17, 2004 We present a novel concept for the creation of lipid vesicle microarrays based on a patterning approach termed Molecular Assembly Patterning by Lift-off (MAPL). A homogeneous MAPL-based single-stranded DNA microarray was converted into a vesicle array by the use of vesicles tagged with complementary DNAs, permitting sequence-specific coupling of vesicles to predefined surface regions through complementary DNA hybridization. In the multistep process utilized to fulfill this achievement, active spots consisting of PLL-g-PEGbiotin with a resistant PLL-g-PEG background, as provided by the MAPL process, was converted into a DNA array by addition of complexes of biotin-terminated DNA and NeutrAvidin. This was then followed by addition of POPC vesicles tagged with complementary cholesterol-terminated DNA, thus providing specific coupling of vesicles to the surface through complementary DNA hybridization. Quartz crystal microbalance with dissipation (QCM-D) and optical waveguide lightmode spectroscopy monitoring were used to optimize the multistep surface modification process. It was found that the amount of adsorbed biotinDNA-NeutrAvidin complexes decreases with increasing molar ratio of biotinDNA to NeutrAvidin and decreasing ionic strength of the buffer solution. Modeling of the QCM-D data showed that the shape of the immobilized vesicles depends on the amount of available anchoring groups between the vesicles and the surface. Fluorescent microscopy images confirmed the possibility to create welldefined patterns of DNA-tagged, fluorescently labeled vesicles in the micrometer range.
1. Introduction The merit of microarray-based systems is that they allow a large number of different biomolecular recognition reactions to be measured within a small area. Since this strongly aids parallel processing of data, this concept has improved both the speed and accuracy by which genetic and proteomic information can be obtained. The success of DNA microarrays1-3 relies on a combination of advancement in PCR technology and the ability of complementary nucleotide strands to hybridize with singlestranded DNA on defined locations on array formats. Although this principle gives information about the active genetic content (i.e., expressed mRNAs), proteins, not genes, are responsible for orchestrating the multitude of still poorly understood biochemical reactions that make life possible. Accordingly, there is an urgent need for analytical concepts that, in analogy with DNA arrays, make high-throughput analysis of proteins possible. Current protein microarrays can, in principle, be divided into two sub areas: one being focused on assays for functional analysis of proteins4 and one on the use of natural protein-capture agents such as antibodies5,6 or aptamers7,8 to identify/quantify the abundance of predefined proteins in complex mixtures. * Author to whom correspondence should be addressed. † Swiss Federal Institute of Technology (ETH) Zurich. ‡ Chalmers University of Technology and Go ¨ teborg University. (1) Sanders, G. H. W.; Manz, A. Chip-based microsystems for genomic and proteomic analysis. TrAC, Trends Anal. Chem. 2000, 19, 364-378. (2) Heller, M. J. DNA microarray technology: Devices, systems, and applications. Annu. Rev. Biomed. Eng. 2002, 4, 129-153. (3) Geschwind, D. H. DNA microarrays: translation of the genome from laboratory to clinic. Lancet Neurol. 2003, 2, 275-282. (4) Troitskaya, L. A.; Kodadek, T. Peptides as modulators of enzymes and regulatory proteins. Methods 2004, 32, 406-415. (5) Kingsmore, S. F.; Patel, D. D. Multiplexed protein profiling on antibody-based microarrays by rolling circle amplification. Curr. Opin. Biotechnol. 2003, 14, 74-81.
In contrast to DNA arrays, the realization of protein microarrays is challenging due to the fact that proteins cannot be multiplied and that proteins, especially transmembrane proteins, are easily denatured upon immobilization onto a solid substrate. This puts strong constraints on both the mode of detection and the environment provided upon immobilization. Supported lipid bilayers9,10 and vesicles11 are, however, suitable environments for the incorporation of transmembrane proteins. The first membrane protein array was only recently realized using microdispersing but having the drawback that printing includes a dry step what might affect the lipids.12 A novel concept introduced by Niemeyer et al. provides a promising approach for the transformation of DNA arrays into arrays of water-soluble proteins.13 Using lipid vesicles modified with single-stranded DNA, an extension of this principle to a lipid vesicle array was (6) Kusnezow, W.; Jacob, A.; Walijew, A.; Diehl, F.; Hoheisel, J. D. Antibody microarrays: An evaluation of production parameters. Proteomics 2003, 3, 254-264. (7) McCauley, T. G.; Hamaguchi, N.; Stanton, M. Aptamer-based biosensor arrays for detection and quantification of biological macromolecules. Anal. Biochem. 2003, 319, 244-250. (8) Hamaguchi, N.; Ellington, A.; Stanton, M. Aptamer beacons for the direct detection of proteins. Anal. Biochem. 2001, 294, 126-131. (9) Graneli, A.; Rydstrom, J.; Kasemo, B.; Hook, F. Formation of supported lipid bilayer membranes on SiO2 from proteoliposomes containing transmembrane proteins. Langmuir 2003, 19, 842-850. (10) Tillman, T. S.; Cascio, M. Effects of membrane lipids on ion channel structure and function. Cell Biochem. Biophys. 2003, 38, 161190. (11) Kahya, N.; Pecheur, E. I.; de Boeij, W. P.; Wiersma, D. A.; Hoekstra, D. Reconstitution of membrane proteins into giant unilamellar vesicles via peptide-induced fusion. Biophys. J. 2001, 81, 1464-1474. (12) Fang, Y.; Frutos, A. G.; Lahiri, J. Membrane protein microarrays. J. Am. Chem. Soc. 2002, 124, 2394-2395. (13) Niemeyer, C. M.; Adler, M.; Pignataro, B.; Lenhert, S.; Gao, S.; Chi, L. F.; Fuchs, H.; Blohm, D. Self-assembly of DNA-streptavidin nanostructures and their use as reagents in immuno-PCR. Nucleic Acids Res. 1999, 27, 4553-4561.
10.1021/la0482305 CCC: $27.50 © 2004 American Chemical Society Published on Web 11/11/2004
Micropatterning of DNA-Tagged Vesicles
recently presented by Yoshina-Ishii et al.14 and by the coauthors15 whereby membrane-bound proteins could also be incorporated. In the latter work, DNA-modified gold spots were surrounded by a planar lipid bilayer formed on SiO2, known as one of the few surface modifications that resist lipid vesicle adsorption.16 Herein, the previous patterning approach is improved by replacing the preferential Au/SiO2 modification scheme used in ref 15 by the Molecular Assembly Patterning by Lift-off (MAPL) technique.17 In brief, this technique uses a simple dip and rinse approach in combination with standard photolithography to create biotinylated poly(L-lysine)-g-poly(ethylene glycol) (PLL-g-PEGbiotin) modified patterns surrounded by a resistant region of PLL-g-PEG,18-20 which has the following advantages. First, the approach permits the use of transparent substrates which makes it compatible with current microarray readers. Second, the PLL-PEG provides a stable background being highly resistant to nonspecific adsorption of both vesicles and proteins. Third, it is possible to produce active spots in the submicrometer range with controlled density of capture agents. Finally, the patterned substrates can be stored under ambient conditions without loss of performance. Complexes of biotin-terminated DNA and NeutrAvidin, preformed in solution, were immobilized to the biotinylated active spots of the MAPL-modified substrates.19,21,22 Vesicles tagged with complementary cholesterol-terminated DNA23 could then be specifically coupled to the surface through the hybridization of the DNA strands (see schematic illustration in Figure 1). To support an optimization of this type of lipid-vesicle array, the surface density of immobilized biotinDNA was investigated as a function of the ionic strength of the buffer and the molar mixing ratio of biotinDNA to NeutrAvidin. The influence from the amount of available binding sites (i) on the hybridization of the cholesterol complementary DNA (cDNA) and (ii) on the subsequent immobilization of the vesicles was monitored using Optical Waveguide Lightmode Spectroscopy (OWLS).24 The influence of the density of the coupling groups between the surface and (14) Yoshina-Ishii, C.; Boxer, S. G. Arrays of mobile tethered vesicles on supported lipid bilayers. J. Am. Chem. Soc. 2003, 125, 3696-3697. (15) Svedhem, S.; Pfeiffer, I.; Larsson, C.; Wingren, C.; Borrebaeck, C.; Hook, F. Patterns of DNA-labeled and scFv-antibody-carrying lipid vesicles directed by material-specific immobilization of DNA and supported lipid bilayer formation on an Au/SiO2 template. Chembiochem 2003, 4, 339-343. (16) Chapman, D. Biomembranes and New Hemocompatible Materials. Langmuir 1993, 9, 39-45. (17) Falconnet, D.; Koenig, A.; Assi, F.; Textor, M. A Combined Photolithographic and Molecular-Assembly Approach to Produce Functional Micropatterns for Applications in the Biosciences. Adv. Funct. Mater. 2004, 14, 749-756. (18) Ruiz-Taylor, L. A.; Martin, T. L.; Zaugg, F. G.; Witte, K.; Indermuhle, P.; Nock, S.; Wagner, P. Monolayers of derivatized poly(L-lysine)-grafted poly(ethylene glycol) on metal oxides as a class of biomolecular interfaces. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 852857. (19) Huang, N. P.; Voros, J.; De Paul, S. M.; Textor, M.; Spencer, N. D. Biotin-derivatized poly(L-lysine)-g-poly(ethylene glycol): A novel polymeric interface for bioaffinity sensing. Langmuir 2002, 18, 220230. (20) Huang, N. P.; Michel, R.; Voros, J.; Textor, M.; Hofer, R.; Rossi, A.; Elbert, D. L.; Hubbell, J. A.; Spencer, N. D. Poly(L-lysine)-g-poly(ethylene glycol) layers on metal oxide surfaces: Surface-analytical characterization and resistance to serum and fibrinogen adsorption. Langmuir 2001, 17, 489-498. (21) Wilchek, M.; Bayer, E. A. Foreword and introduction to the book (strept)avidin-biotin system. Biomol. Eng. 1999, 16, 1-4. (22) Rosano, C.; Arosio, P.; Bolognesi, M. The X-ray three-dimensional structure of avidin. Biomol. Eng. 1999, 16, 5-12. (23) Ohvo-Rekila, H.; Ramstedt, B.; Leppimaki, P.; Slotte, J. P. Cholesterol interactions with phospholipids in membranes. Prog. Lipid Res. 2002, 41, 66-97.
Langmuir, Vol. 20, No. 26, 2004 11349
Figure 1. Schematics of the multistep surface-modification process for the immobilization of intact vesicles using DNA as coupling agent. A Nb2O5-coated substrate is modified with PLL-g-PEG and PLL-g-PEGbiotin. BiotinDNA-NeutrAvidin complexes bind to the immobilized biotin on the surface. Complexes of cholesterol-tagged complementary DNAs and POPC vesicles are hybridized to the immobilized biotinDNA.
the vesicles was further investigated using the Quartz Crystal Microbalance with Dissipation (QCM-D) technique.25 2. Materials and Methods 2.1. Materials. All measurements were carried out in either a 10-mM 4-(2-hydroxyethyl)piperazine-1-ethane-sulfonic acid (HEPES 1) (MicroSelect, Fluka Chemie GmbH, Switzerland) solution or a 160-mM solution, consisting of 10-mM 4-(2-hydroxyethyl)piperazine-1-ethane-sulfonic acid and 150-mM NaCl (MicroSelect, Fluka Chemie GmbH, Switzerland) (HEPES 2). The pH of both buffer solutions was adjusted to pH ) 7.4. Both buffers were made with ultrapure water, (Milli-Q gradient A 10 system, resistance ≈ 18 MΩ cm, TOC < 4 ppb, Millipore Corporation, USA), stored at 4 °C and filtered with a 0.2-µm pore filter (Sigma Aldrich Chemie GmbH, Germany) prior to use. PLL-g-PEG, a graft copolymer with a poly(L-lysine) (PLL) backbone of 20 kDa and poly(ethylene glycol) (PEG) side-chains of 2 kDa, and a grafting ratio of 3.5, was used for all experiments. The biotinylated polymer (PLL-g-PEGbiotin) had the same architecture, with 50% of its side-chains biotinylated using PEGbiotin and a molecular weight of 3.4 kDa. Both polymers were synthesized and characterized as previously described in detail.19 DNA strands (5′-TAG-TTG-TGA-CGT-ACA-CCC-CC-3′, MW ) 6621 g/mol) with 3′ modified with biotin and the complementary 20mer (5′-TGT-ACG-TCA-CAA-CTA-CCC-CC-3′, MW ) 6738 g/mol) with 3′ modified with tri(ethylene glycol) (TEG) cholesterol (MedProbe, Norway) were used. Stock solutions of DNA conjugates (0.7 nmol in 20 µL of Milli-Q water), NeutrAvidin (MW ) 60 000 g/mol, Molecular Probes, The Netherlands) (20 µg in 20 µL of buffer) and streptAvidin Alexa Fluor633 (Molecular Probes, The Netherlands) (50 µg in 20 µL of buffer) were aliquoted and stored at -20 °C. 1-Palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine lipids (POPC; Avanti Polar Lipids, USA) dissolved in chloroform were stored at -20 °C. For fluorescent vesicles, 1% (w/w) of 2-(12-(7-nitrobenz2-oxa-1,3-diyzol-4-yl)amino)dodecanoyl-1-hexadecanoyl-sn-glycero-3-phosphocholine (NBD-HPC; Molecular Probes, USA) was added to the lipid solution. Unilaminar lipid vesicles (stored at 4 °C under N2 atmosphere) were prepared by evaporation of the solvent under Argon (∼1 h) followed by hydration in buffer (5 mg/mL) and extrusion through 50-nm filters (31 times). Commercially available planar optical waveguides (2400 mV, MicroVacuum Ltd., Hungary), QCM sensor crystals (Q-sense AB, Sweden), and Pyrex wafers (Sensorprep Services, Inc., USA) were used for the experiments. An 8-nm thin film of Nb2O5 was sputter coated on all substrates according to already published (24) Voros, J.; Ramsden, J. J.; Csucs, G.; Szendro, I.; De Paul, S. M.; Textor, M.; Spencer, N. D. Optical grating coupler biosensors. Biomaterials 2002, 23, 3699-3710. (25) Marx, K. A. Quartz crystal microbalance: A useful tool for studying thin polymer films and complex biomolecular systems at the solution-surface interface. Biomacromolecules 2003, 4, 1099-1120.
11350
Langmuir, Vol. 20, No. 26, 2004
protocols.26 The following cleaning protocols were applied to the waveguides and to the QCM sensor crystals, respectively, prior to use: (i) 10 min ultrasonication in 2-propanol (spectroscopy grade, Fluka Chemie GmbH, Switzerland), 10 min ultrasonication in Milli-Q water, rinsing with Milli-Q water and blow-drying with nitrogen, and 1-min oxygen plasma cleaning and (ii) soaking overnight in 2% lauryl sulfate (Sigma Aldrich Chemie GmbH, Germany), rinsing with Milli-Q water, and UV/ozone cleaning for 30 min. 2.2. Optical Waveguide Lightmode Spectroscopy (OWLS). OWLS experiments were carried out in BIOS-1 instruments (Artificial Sensing Instruments, Switzerland) using a laminar flow-through cell (8 × 2 × 1 mm3). De Feijter’s formula was applied to calculate the adsorbed mass.27 The refractive index increment (dn/dc) of 0.13 cm3/g for PLL-g-PEG-based polymers and 0.182 cm3/g for proteins and vesicles was used for subsequent calculations of adsorbed mass.19 Since no dn/dc value for vesicles is available in the literature, a value of dn/dc ) 0.182 cm3/g was only applied for qualitative comparison. Waveguides were initially placed in buffer (HEPES 1 or 2) immediately following the cleaning and allowed to soak overnight. The samples were exposed, in situ, to the PLL-g-PEGbiotin solution in a concentration of 0.1 mg/mL. The adsorption was subsequently monitored until the surface was saturated. Then, the polymer solution was replaced with buffer. Next, the PLL-g-PEGbiotin-modified samples were exposed to solutions consisting of different molar concentration ratios of biotinDNA to NeutrAvidin (Rb) dissolved in either HEPES 1 or HEPES 2 buffer. Rb was varied between Rb ) 0 and 6 using a constant NeutrAvidin concentration (20 µg/mL) by changing the concentration of biotinDNA. The NeutrAvidin and the biotinDNA solutions were mixed together (v:v ) 1:1) and, after 10 min of incubation, injected into the flow cell of the OWLS instrument. After the surface was saturated, the mixture was replaced with buffer. Then, the modified waveguide was exposed in situ to a cholesterol cDNA solution in a concentration of 0.7 µM until saturation and then subsequently rinsed with buffer. The nonspecific adsorption of POPC vesicles in HEPES 2 was shown at different stages of the experiment: first, on PLL-g-PEG, second, on PLL-g-PEGbiotin, third, on NeutrAvidin, and fourth, on adsorbed biotinDNA-NeutrAvidin complexes. The experiment was stopped at these four different stages, and the POPC vesicles (0.25 mg/mL) were injected. After 10 min, the vesicle solution was replaced with buffer and the amount of nonspecifically adsorbed vesicles was recorded. 2.3. Quartz Crystal Microbalance with Dissipation (QCM-D). The QCM-D instrument (Q-Sense AB, Sweden) measures changes in the frequency (f) and dissipation factor (D) of an oscillating quartz crystal upon adsorption of a viscoelastic layer.26,28,29 Immediately after cleaning, the crystal was mounted into the liquid-exchange cell of the instrument. The solution and the cell were temperature-stabilized at 21.5 ( 0.03 °C. Cholesterol cDNA-POPC vesicle complexes were hybridized to biotinDNA-modified crystals using different surface densities of biotinDNA (Rb) and changing the molar ratio of cholesterol cDNA to POPC vesicles in solution (Rc). Rc was varied between Rc ) 0.5 and 13 using a constant POPC vesicles concentration (0.25 mg/mL) by changing the concentration of the cholesterol cDNA. The two solutions were mixed together (v:v ) 1:1) and, after 10 min of incubation, the biotinDNAmodified crystal (as described in section 2.2) was exposed in situ to the solution. (26) Hook, F.; Voros, J.; Rodahl, M.; Kurrat, R.; Boni, P.; Ramsden, J. J.; Textor, M.; Spencer, N. D.; Tengvall, P.; Gold, J.; Kasemo, B. A comparative study of protein adsorption on titanium oxide surfaces using in situ ellipsometry, optical waveguide lightmode spectroscopy, and quartz crystal microbalance/dissipation. Colloids Surf., B 2002, 24, 155-170. (27) Defeijter, J. A.; Benjamins, J.; Veer, F. A. Ellipsometry as a Tool to Study Adsorption Behavior of Synthetic and Biopolymers at AirWater Interface. Biopolymers 1978, 17, 1759-1772. (28) Keller, C. A.; Kasemo, B. Surface specific kinetics of lipid vesicle adsorption measured with a quartz crystal microbalance. Biophys. J. 1998, 75, 1397-1402. (29) Hook, F.; Rodahl, M.; Brzezinski, P.; Kasemo, B. Energy dissipation kinetics for protein and antibody-antigen adsorption under shear oscillation on a quartz crystal microbalance. Langmuir 1998, 14, 729-734.
Sta¨ dler et al.
Figure 2. A sample curve of a cholesterol cDNA immobilization experiment monitored with OWLS in HEPES 2. First, PLL-g-PEGbiotin is adsorbed on a Nb2O5-coated waveguide (step 1). Next, the biotinDNA-NeutrAvidin complexes are immobilized (step 2). The cholesterol cDNA is hybridized to the immobilized biotinDNA (step 3a). To investigate the possibility to strengthen the DNA-mediated interaction between the immobilized vesicles and the surface, the vesicle-modified QCM crystal was exposed in situ to a solution containing further cholesterol cDNA molecules in a concentration of 0.35 µM. 2.4. Confocal Laser Scanning Microscope (CLSM). All fluorescent experiments were performed on a confocal laser scanning microscope (Confocal ZeissLSM 510, Germany) using a 10×, 0.3 NA objective (Zeiss, Germany). The MAPL process was applied to Nb2O5-coated glass samples as described by Falconnet et al. in detail.17 Briefly, photolithography was used to create substrates with a photoresist-protected background and unprotected areas. Subsequently, PLL-g-PEGbiotin was immobilized on the bare oxide areas by spontaneous assembly from aqueous solution. The photoresist was then lifted off in 1-methyl-2-pyrrolidone (for peptide synthesis, Fluka Chemie GmbH, Switzerland), followed by backfilling the background with PLL-g-PEG. The patterned samples (PLL-g-PEGbiotin squares of 200 × 200 µm2 separated by a distance of 300 µm in a PLL-g-PEG background) were mounted into a poly(dimethylsiloxane) (PDMS) flow cell which was filled with buffer solution. BiotinDNAstreptAvidin Alexa Fluor633 complexes were introduced in a molar ratio of Rb ) 2. The labeled streptAvidin was used in order to visualize the patterns. For the immobilization of the vesicles, a cholesterol cDNA-POPC/NBD-HPC vesicle mixture (Rc ) 13) was introduced. After a 40 min incubation, the mixture was replaced with buffer and the patterns were observed with the CLSM.
3. Results In this section, we present the results of the OWLS and QCM-D measurements performed on homogeneous substrates following the order of the multistep surface modification process depicted in Figure 1. While OLWS provides detailed information about the adsorbed molecular mass, QCM-D provides further information about the viscoelastic properties of the vesicles upon immobilization. Finally, the data obtained from the QCM-D and the OWLS analysis were used to determine appropriate conditions for the confocal laser scanning microscopy images of patterned vesicles on MAPL-modified surfaces. 3.1. Multistep Surface Modification. The adsorption of PLL-g-PEGbiotin onto a Nb2O5-coated surface (step 1) and the immobilization of biotinDNA-NeutrAvidin complexes (step 2) in HEPES 2 were monitored with OWLS and QCM-D (Figure 2 and Figure 5). PLL-g-PEGbiotin forms a full monolayer (218 ( 16 ng/cm2) within 40 min, as expected form previous studies.30,31 The complexes of biotinDNA-NeutrAvidin saturate the surface within 1 h (30) Kenausis, G. L.; Voros, J.; Elbert, D. L.; Huang, N. P.; Hofer, R.; Ruiz-Taylor, L.; Textor, M.; Hubbell, J. A.; Spencer, N. D. Poly(L-lysine)g-poly(ethylene glycol) layers on metal oxide surfaces: Attachment mechanism and effects of polymer architecture on resistance to protein adsorption. J. Phys. Chem. B 2000, 104, 3298-3309.
Micropatterning of DNA-Tagged Vesicles
Figure 3. The OWLS results of the influence of the ionic strength of the buffer on the amount of adsorbed biotinDNANeutrAvidin complexes as a function of the molar ratio of biotinDNA to NeutrAvidin in solution (Rb). (See steps 1 and 2 in Figure 2). An ionic-strength-dependent decay of the adsorbed mass of biotinDNA-NeutrAvidin complexes with increasing Rb is observed.
Langmuir, Vol. 20, No. 26, 2004 11351
Figure 4. The calculated mean distance between the immobilized cholesterol cDNA molecules as a function of the molar ratio of biotinDNA to NeutrAvidin in solution (Rb). (Note the logarithmic scale for Rb. * denotes linear interpolated values.) A minimum in the distance is observed for Rb ) 2 corresponding to 6.7 nm. The original OWLS experiments, which were used to calculate the mean distances, are summarized in the insert which shows the determined amounts of cholesterol cDNA (0.7 nmol/mL) hybridized to the immobilized biotinDNA as a function of Rb in HEPES 2. (See step 3a in Figure 2.) Rb ) 2 leads to the highest amount of hybridized cDNA while, for Rb ) 0.01, the adsorbed amount was lower than the detection limit of the instrument (
