|Home | About | Journals | Submit | Contact Us | Français|
The fabrication of single virus arrays is herein demonstrated using the direct printing of unmodified anti-M13 bacteriophage antibodies on silicon with nanometer resolution, widely variable feature pitch, and flow alignment of the viruses. Organization of virus-based systems into functional, addressable arrays has many technological applications, including micro-array technology and bottom-up nano-assemblies.
Many self-assembly schemes based on those found in biological systems have been demonstrated for the organization of inorganic and biological materials on the nanometer scale. Extending these examples into practical use relies on the ability to hierarchically organize them over arbitrary length scales. Achieving this level of control, however, has been hindered by the incompatibility of biological materials with current processing methods. Efforts for bridging this gap have mostly focused either on nonspecific chemical modification of surfaces, alteration of the naturally occurring system, or a combination thereof. The desire exists to develop general biocompatible processes for the organization of unmodified biological systems that capitalize on the numerous highly specific interactions commonly found in nature, including DNA, antibodies, and protein complexes. To this end, the fabrication of single virus arrays is herein demonstrated using the direct printing of unmodified anti-M13 bacteriophage antibodies on silicon with nanometer resolution and widely variable feature pitch.
The intersection of biology and technology has provided many unique solutions to challenges in both fields. Technological advances have allowed biological systems to be studied with ever-increasing detail and reproducibility. Alternately, biologically inspired approaches have shown great promise for the self-assembly and directed assembly of materials on the nanometer scale. The filamentous M13 bacteriophage virus has exhibited a tremendous capacity for incorporating biological and inorganic materials (including metallic, magnetic, and semi-conducting materials) into its self-assembled, genetically modifiable architecture. Macroscopic organization of M13 bacteriophages has been achieved using liquid crystalline behavior, phase separation phenomena, and virus-membrane complexes to create materials of high uniformity and element density. Nevertheless, these methods are not applicable for the fabrication of addressable arrays of single elements. Methods for patterning viruses, including chemical linkers, nucleic acid hybridization, and metal ions have been demonstrated, but often face a tradeoff between specificity and generality of the approach. The use of highly specific antibody interactions, however, has remained relatively unexplored. This has mainly been due to the gross loss of antibody activity during sample preparation and processing. Soft lithographic methods, such as microcontact printing, have been successful in maintaining biomolecular activities, but remain challenged by the vast range of length scales on which biological interactions occur: proteins and viruses (nanometer), cells (micrometer), and tissue (millimeter). This limitation in feature size and pitch is due to the mechanical properties of the elastomeric materials used in the printing of proteins, mainly polydimethylsiloxane (PDMS). To overcome this limitation, a subtractive printing technique has recently been developed as a versatile method for the patterned transfer of antibodies from solution to substrate through a series of step-wise reductions in nonspecific hydrophobic interactions (Fig. 1). This method benefits from the use of a featureless elastomer enabling feature sizes, pitches, and total patterned areas that are independent of its mechanical properties.[14,15] These parameters are therefore defined by the lithographic process used in fabricating the template master (Fig. 1b). A judicious choice of substrate and elastomeric materials allows for the direct transfer of biological material without the need for chemical modification of either the substrate or the biological system. Herein, we apply the subtractive contact printing technique for the nanometer-scale patterning of antibodies with micrometer pitch to capture individual M13 bacteriophages. Further, we explore the effects of both the solution parameters and antibody feature size for the optimization of phage-pattern interactions.
The complexity of biological systems creates large interdependencies on pH, ionic valency and strength, and concentration, which can greatly complicate the driving forces governing immobilization of biological entities to surfaces. M13 bacteriophage solutions undergo radical physical transformations under minor solution variations due to the filamentous structure (880 × 6 nm2) and large negative surface charge density (SCD, σ) of the virus, which is a known function of pH (σM13 = 1e−/256 A2 for pH ≥ 7; for comparison, σDNA = 1e−/106 A2).[18,19] Therefore solution conditions were optimized for the binding of M13 bacteriophage to macroscopic antibody patterns (2 × 2 μm2 features) to decouple these effects when studying the impact of the feature size. Maintaining a large negative SCD during phage binding was necessary to minimize multiple-site occupancy and nonspecific background binding by increasing phage-phage and phage-silicon electrostatic repulsion, as silicon also has a negative SCD under standard buffer conditions (1e−/2381 A2). Reduction of the ionic strength of the buffered phage solutions by 50% (< 75 mM NaCl) was used to minimize charge-screening effects. Optimization of the binding conditions resulted in complete coverage of the patterned antibody, with only minimal nonspecific background binding to the substrate. Importantly, for the given system, where a repulsive electrostatic interaction exists between the phage and substrate, there was no need for surface passivation.
Achieving single-element arrays requires controlling both the antibody feature size and the binding kinetics. Although reducing the solution concentration of phage can be used to statistically achieve single-element site occupancy, this approach is limited by the binding affinity of the capture antibody. The lower detection limit of the printed anti-M13 bacteriophage capture antibody was determined using patterns having average feature sizes of 240 × 240 nm2, and was consistent with the supplier-recommended working dilution of 107 plaque forming units (pfu)/mL. Phage solutions in the range 107 to 109 pfu/ml incubated with the antibody patterns produced individual, well-separated immobilized phages with increasing site occupancy and pattern coverage (Fig. 2). At phage concentrations above 109 pfu/mL, local changes in the binding statistics were observed and are suggestive of large inhomogeneities in the phage solution. At the highest concentrations studied (1010–1011 pfu/mL), changes in the interactions resulted in phage bundling and the creation of star-like patterns.
Understanding the interactions between the bacteriophage protein coat and the patterned antibody is necessary for achieving single site occupancy. On the 2 × 2 μm2 macroscopic patterns, atomic force microscopy (AFM) analyses revealed two bacteriophage binding conformations, in which either complete immobilization of the protein coat or localization to the feature edge occurs. Decreasing the feature size below ~625 nm promoted the predominantly edge-binding regime as a result of the physical size and persistence length of the M13 bacteriophage (the commonly reported value is 2.2 μm, with recent reports suggesting an even shorter length of 1.2 μm). The extension of the phage off of the antibody feature increases the repulsive electrostatic phage-silicon interaction, driving the majority of the protein coat into solution. Patterns having average antibody features of 240 × 240 nm2, 200 × 200 nm2, and 90 × 90 nm2 were incubated with a phage solution of 109 pfu/mL at a pH and ionic strength as previously optimized. Antibody patterns having average feature sizes of 240 × 240 nm2 had a majority of sites occupied by two or more phages (this was found to be easily manipulable by minor changes in solution conditions). Reduction of the antibody feature size to 200 × 200 nm2 achieved arrays with 42% single site occupancy and high coverage with a greater degree of reproducibility than the larger patterns had. However, a number of sites having two or more phages (21%) still remained. Further reduction of the antibody feature size to 90 × 90 nm2 achieved complete single site occupancy at the cost of low coverage (20%). At these dimensions, the low occupancy probably originates from the detachment of the phage-bound antibodies during sample washing. The high aspect ratio of the M13 bacteriophage provides a sufficiently large hydrodynamic coefficient of drag for alignment in fluid flows. Given the extent of the phage coat in solution for the nanoscale features, control over the direction of arrayed phage was achievable by using flow alignment. This enabled a four-fold increase in the density of the arrayed phage by decreasing the interfeature spacing from 2.5 to 1.0 μm (Fig. 3). Multiple fields with antibody islands having a size of 200 × 200 nm2 and with an area of 0.25 mm2 were patterned in one step on silicon substrates. These fields can be repeated over a total area of 30 × 30 mm2 using reasonable (24 h or less) e-beam writing times. Given a 1 μm pitch between islands and an average phage occupancy per island of 42%, at least 3 × 105 phages can be arrayed per mm2 of substrate. In practice, a phage library having a concentration of 109 pfu/mL can easily be made to a volume of 100 μL, which is sufficient to incubate several cm2 of substrate. The maximum size of an array is therefore not limited by the antibody patterning or phage assembly steps. Observation of extensive bending of the phage in the liquid flow implies a strong antibody-protein binding and suggests a possible means of studying the persistence length of filamentous systems. Increasing the phage density and alignment to prefabricated structures for the creation of more complex architectures can therefore be realized using the combination of subtractive printing and flow alignment.
We extended the previous experiments, which aimed at identifying parameters responsible for non-specific phage deposition, phage bundling, and island occupancy, in order to refine the conditions for which at least one phage is present per island, (Fig. 4). Islands (n = 114) with a lateral size of ~250 nm were analyzed using atomic force microscopy and a strong reduction in the chance of having at least one phage per island was observed only when the concentration decreased below 5 × 109 pfu/mL. In comparison, 100 nm islands (n = 299) had a strongly reduced fractional occupancy: the fractional occupancy was only 59% at a starting concentration of 5 × 1010 pfu/mL and diminished to 18% and 3% for phage concentrations of respectively 5 × 109 and 5 × 108 pfu/mL. Two sets of conditions may therefore be used for arraying phages. If single phage arrays are desired, they should be arrayed on 100 nm antibody island at a concentration of 5 × 1010 pfu/mL. If arrays with very high density are desired and multiple island occupancy is unimportant for example if identical phages must be arrayed, then 250 nm islands and a phage concentration of 5 × 1010 pfu/mL should be selected.
Using nanoscale patterns of antibodies directed against a phage coat protein as it was done here provides a general strategy for arraying phages irrespectively of their biological diversity: identical phages or different phages forming a library can be arrayed in the same way. The arrayed phages may subsequently be exposed to ligands of interest and phage-ligand interactions may be identified using simple surface fluorescence assays. Organization of biological systems into functional, addressable arrays has many technological applications, including micro-array technology and bottom-up nano-assemblies. Beyond the technical implications, addressable arrays of individual biological components have the potential to elucidate the intricate relationships between spatial organization and resulting functionality in biological systems. Macroscopic cellular activities such as proliferation, migration, and differentiation all rely on interactions with elements whose size and organization are defined at the nanoscale. Extending the understanding of these processes from the ensemble to the molecular level will enable more advanced diagnostics and therapeutics.
High-resolution nanotemplates were produced using electron-beam lithography. PMMA-resist-coated silicon wafers were exposed in an e-LiNE electron-beam lithography system (voltage: 20 kV, aperture: 10 μm, beam current: 29 pA) (Raith GmbH, Dortmund, Germany), developed in a solution of MIBK:isopropanol at a 1:3 ratio for 30 s, immersed in isopropanol for 1 min, and blown dry under a stream of N2. The PMMA pattern was transferred into the silicon substrate using a low-etch-rate reactive ion etcher in a balanced process that used SF6 as precursor for the etching and C4F8 for passivation of the sidewalls (Alcatel Vacuum Technology France, Annecy, France), which lasted for 25 s.
Sylgard® (Dow Corning, Midland, MI) 184 PDMS elastomers were cured at 60°C for at least 24 h in Petri dishes. The side of the elastomer that was in contact with the Petri dish was inked with ~100 μL of antibody solution for 45 min. Anti-fd Bacteriophage (B7786, Sigma, St. Louis, MO) was used at a concentration of 0.1 mg/mL in phosphate-buffered saline (PBS) (A7906, Sigma). After inking, elastomers were rinsed using PBS and deionized water, and blown dry under a stream of N2 for ~30 s.
Details of the subtractive printing technique have been previously published. Briefly, clean silicon substrates and nanotemplates were treated with oxygen plasma at 200 W for 60 s (Technics Plasma 100-E, Florence, KY). Proteins on homogenously inked elastomers were removed in selected areas by bringing the elastomers into contact with the nanotemplate for 15 s, followed by manual release. The protein patterns were transferred from the elastomers to the final substrates using a 30-second-long printing step. Intimate contact between the elastomer and the nanotemplate/substrate occurred after placing the elastomer on the nanotemplate/substrate by hand and applying a slight pressure with tweezers. Before reuse, the nanotemplates were cleaned of organic material by repeating the treatment with oxygen plasma.
Atomic force microscopy (AFM) images were obtained using a Nanoscope Dimension 3000 (Digital Instruments, Santa Barbara, CA) operated in tapping mode using standard silicon cantilevers (174–191 kHz, Nanosensors, Neuchâtel, Switzerland). AFM images were planarized, displayed, and analyzed using NanoScope 6.12r1 software.
M13 bacteriophage stock (New England Biolabs) was amplified in the host bacteria E.coli (ER2738 NEB) using standard phage methods. Briefly, phage stock (1 × 1012 pfu/mL) was added to a 1:100 dilution of an overnight culture of bacteria and incubated with shaking at 37°C for 5.5 h. Phages were separated from bacteria via centrifugation and concentrated by polyethylene glycol/NaCl precipitation overnight at 4°C, followed by centrifugation. Dialysis of the resulting phage was used to remove excess salts and ensure proper pH.
5 mL of phage stock in TBST, Tris-Buffered Saline (TBS) plus 0.1% Tween-20 (Sigma Aldrich) was incubated under gentle agitation (without using convective flow) with the subtractively printed substrates for 1 h, followed by gentle, but thorough, washing using TBST, TBS, water (18.2 M⃠) and dried with compressed nitrogen. Samples were placed in a vacuum desiccator overnight prior to AFM analysis. In some experiments, viruses were aligned by rinsing them in one direction after the immobilization step and before drying the sample.
We thank H. Riel, E. Lörtscher, T. Kraus and U. Drechsler for their support with the fabrication of nanotemplates, and H. Wolf and M. Zimmermann for discussions. E.D. acknowledges partial support of the State Secretariat for Education and Research (SEC) in the framework of the EC-funded project NaPa (Contract No. NMP4-CT-2003-500120). Partial funding was provided by National Institutes of Health (R01-GM065918 to A.J.G.) and the Whitaker International Fellows and Scholars Program (to S.R.C.). We also thank W. Riess and P. Seidler for their continuous support.
Daniel J. Solis, IBM Research GmbH, Zurich Research Laboratory, 8803 Rüschlikon (Switzerland)
Sean R. Coyer, IBM Research GmbH, Zurich Research Laboratory, 8803 Rüschlikon (Switzerland). Woodruff School of Mechanical Engineering, Petit Institute for Bioengineering, and Bioscience, Georgia Institute of Technology, Atlanta, GA 30332-0363 (USA)
Andrés J. García, Woodruff School of Mechanical Engineering, Petit Institute for Bioengineering, and Bioscience, Georgia Institute of Technology, Atlanta, GA 30332-0363 (USA)
Emmanuel Delamarche, IBM Research GmbH, Zurich Research Laboratory, 8803 Rüschlikon (Switzerland)