Biomacromolecules 2008, 9, 355–362
355
Visual Detection of Labeled Oligonucleotides Using Visible-Light-Polymerization-Based Amplification Ryan R. Hansen, Hadley D. Sikes, and Christopher N. Bowman* Department of Chemical and Biological Engineering, ECCH 111 CB 424, University of Colorado, Boulder, Colorado 80309 Received June 15, 2007; Revised Manuscript Received September 23, 2007
DNA biochip technology holds potential for highly parallel, rapid, and sensitive genetic diagnostic screening of target pathogens and disease biomarkers. A primary limitation involves a simultaneous, sequence-specific identification of low copy number target polynucleotides using a clinically appropriate detection methodology that implements only inexpensive detection instrumentation. Here, a rapid (20 min), nonenzymatic method of signal amplification utilizing surface-initiated photopolymerization is presented in glass microarray format. Visible light photoinitiators covalently coupled to streptavidin were used to bind biotin-labeled capture sequences. Amplification was achieved through subsequent contact with a monomer solution and the appropriate light exposure to generate 20-240-nm-thick hydrogel layers exclusively from spots containing the biotin-labeled DNA. An amplification factor of 106 to 107 was observed as well as a detectable response generated from as low as ∼104 labeled oligonucleotides using minimal instrumentation, such as an optical microscope or CCD camera.
Introduction The development of molecular diagnostic technology for the detection of infectious disease has become an essential requirement for health improvement, specifically in low-income regions and developing countries.1 Biochips hold vast potential in diagnostic applications due to their specific, parallel detection capability of nucleic acid hybridizations. As subattomole quantities of targeted genetic material are available from a patient sample, on-chip signal amplification methods become a fundamental prerequisite for developing highly sensitive detection methods for targeted pathogens or genetic biomarkers. Enzymatic bioamplification such as polymerase chain reaction (PCR) based methods,2 rolling circle amplification,3 dendrimeric DNA,4 or catalyzed reporter deposition methods,5 coupled with avidin–biotin complex systems6 provides increased sensitivity and specificity at the expense of cost and complexity. Recently, enzyme-free signal amplification methods such as nanoparticle-based bioamplification7–10 have also been developed, as well as label-free detection methods.11,12 However, many of these approaches still rely on sophisticated detection instrumentation not feasible in point-of-care facilities. Sikes et al. recently reported a novel, visual detection method utilizing photopolymerization for signal amplification13 on a commercially available biodetection platform.14,15 This result was achieved by coupling water-soluble Irgacure 2959 photoinitiatiors and streptavidin proteins to a high-molecular-weight copolymer backbone using carbodiimide coupling chemistry,16 then incubation of the biotinylated biochip with this macrophotoinitiator. Contact with a hydroxyethyl acrylate and ethylene glycol dimethacrylate monomer mixture and subsequent exposure to 365 nm light resulted in a macroscopically observable response down to ∼1000 molecular recognition events. While this method showed remarkable sensitivity and amplification, analysis of these surfaces was limited to visual inspection due to the fact that the resulting hydrogel underwent excessive swelling and damage on the surface with rinsing. This system * Corresponding author e-mail:
[email protected].
also has the disadvantages of requiring the use of ultraviolet radiation, potentially causing unwanted, nonspecific bulk polymerization, and also requiring the use of often toxic monomer reagents. An ideal biodection assay would use innocuous reagents and processing conditions while rapidly producing a highly detectable response that is specific, stable, and quantitative. To develop this amplification technology further toward these standards, new initiation systems and monomer formulations were investigated on general glass microarray surfaces compatible with standard microarray scanners or imagers. The use of visible light sources for photoinitiation has the attractive characteristic of requiring only a low-power, inexpensive, and mild excitation source. Further, the use of visible light has the added advantage of eliminating the unwanted bulk polymerization often observed when using UV light. This allows for the selection of a wider variety of monomer formulations that were incompatible with the previous system, particularly formulations containing high concentrations of bifunctional monomers that form thick, highly cross-linked polymers that remain stable on the surface with rinsing. The formation of a surface-stable hydrogel over labeled DNA allows us to fully characterize the amplification process with film thickness and spectroscopic measurements. Finally, visible light enables more efficient amplification due to its higher penetration capability in UV absorbent monomer formulations containing fluorescent monomers or on UV absorbent surfaces characteristic of glass biochips containing surface-bound biomolecules. In this article, we extend this polymerization-based amplification method to visual detection of biotin-labeled DNA functionalized on glass microarray surfaces with the use of photoreducible dyes which initiate upon visible light exposure. We chose eosin isothiocyanate (EITC) as our photoinitiator due to its favorable absorption characteristics and demonstrated its ability to transduce biomolecular recognition into a macroscopically observable response in a highly sensitive manner. A tertiary amine co-initiator is added to the bulk monomer to generate free radicals for the polymerization of a polyethylene glycol
10.1021/bm700672z CCC: $40.75 2008 American Chemical Society Published on Web 12/01/2007
356 Biomacromolecules, Vol. 9, No. 1, 2008
Hansen et al.
Scheme 1
diacrylate (PEGDA) monomer solution. Similar systems have previously been reported for surface-initiated polymerization from silica nanoparticles,17 from aminosilinated glass surfaces for photopatterning applications,18–20 and in cell encapsulation studies.21 The demonstration of DNA detection from polymerization-based signal amplification on general microarray surfaces enables its use as a detection method in applications commonly found in microarray technology, such as single nucleotide polymorphism screening22,23 and should be directly applicable to implementation on commercially available microarray platforms.24
Experimental Section Microarray Fabrication. Commercially available amino or aldehyde functionalized glass substrates were purchased from CEL Associates, and all slides were stored in a vacuum at room temperature. The surface was spotted through pin deposition using a solid pin to deposit ∼570µm-diameter spots and a quill pin for ∼100-µm-diameter spots at 75% humidity. An oligonucleotide sequence (5′ amino-CATCACACAACATCACACAACATCACGTATATAAAACGGAACGTCGAAGG-3′ TEG biotin; Operon) was spotted at an overall concentration of 20 µM on aldehyde substrates in the spotting buffer (3× SSC, 0.05% SDS) using a VersArray ChipWriter Pro system made by Bio-Rad Laboratories. The identical, unlabeled capture sequence was spotted on the surface at a concentration of 4 µM, which has been shown to be optimal for hybridizations. These slides were left in humidity for 24 h. Aminosilane slides were spotted with 4 µM concentrations of 5′ biotin, 5′ Cy3, or unlabeled versions of this same sequence with varied concentrations of labeled sequences present for the fabrication of dilution chips. These were left in humidity for 30 min, then dried in an oven at 80 °C for 2 h, and finally cross-linked to the surface using a 254 nm light. All spotted slides were stored at -18 °C until use. Photoinitiator Synthesis. The visible light photoinitiator eosin-5isothiocyanate (Invitrogen) was functionalized directly onto external lysine residues of streptavidin through the formation of a thiourea bond25 according to the reaction in Scheme 1. Streptavidin was dissolved in a carbonate buffer (0.10 M NaCO3, pH 9) at a concentration of 10 mg/mL. A 10 mg/mL solution of EITC in DMSO was prepared and immediately added to streptavidin at a volumetric ratio of 1:10. The solution was reacted for 8 h at 4 °C, then diluted to a streptavidin concentration of 1 mg/mL in 1× PBS and purified using gel filtration. The product was characterized with conventional UV–vis spectroscopy, and the characteristic peak from EITC at 525 nm was compared to the characteristic protein peak at 280 nm according to eq 1:
nEITC ⁄ nSA )
AbsEITC,525 ⁄ εEITC,525 AbsSA,280 - AbsEITC,280 ⁄ εSA,280
(1)
When 0.2 mg/mL of solution of the product diluted in 1× PBS buffer was used, an average photoinitiator-to-protein ratio of 2.3 was observed. The product was stored at 4 °C and protected from light exposure until further use. Microarray Functionalization of Photoinitiator Product. Spotted slides were blocked with 2 wt % dry milk in ddH20 for 2 h to prevent
nonspecific adsorption of the photoinitiator product to the surface. Slides were then rinsed with water and contacted with 200 µL at 1 µg/mL of visible photoinitiator product in 1× PBS and 5× Denharts solution for 30 min. Slides functionalized with streptavidin-EITC were either placed in boiling water for 2 min or washed in TNT solution (1 M NaCl, 0.1 M Tris, 0.1 wt % Tween 20) to remove nonspecific protein adsorbed on the surface. Slides were rinsed in ddH20 and allowed to dry. Surface-Initiated Photopolymerization. The MEHQ inhibitor was removed from PEGDA (Mn ∼ 575 Da) with the use of a dehibit column (Sigma). Methyl diethanol amine (MDEA) and 1-vinyl-2-pyrolidinone were used as received. The reactive monomer solution consisted of 25 wt % PEG(575)DA, 225 mM MDEA, and 37 mM 1-vinyl-pyrolidinone in a ddH2O solvent. A total of 300 µL of one of the monomer solutions was purged with argon, then contacted with the surface using a Whatman Chip Clip. Slides were placed in an argon chamber that was continuously purged for 5 min prior to radiation. A 400–500 nm, a Novacure collimated light source was used to radiate the streptavidinEITC functionalized slides at 8 mW/cm2 for 20 min. The samples were gently rinsed with ddH20 to remove the unreacted monomer and then dried with argon. Post-Polymerization Surface Characterization of Microarray Surface. A surface profilometer (Dektak 6M) was used to obtain height profiles of the films that were visible on the array after drying the chips such that the hydrogels were completely dehydrated. Stylus force was set at a minimum (1 mg) to minimize mechanical deformation of the hydrogel layers from the 12.5 µm diamond stylus tip. An infrared microscope was used to obtain IR spectra of surface-bound moieties. An Agilent Technologies DNA Microarray Scanner was calibrated at specific settings (100% PMT Gain) with a Cy3 calibration chip obtained from Full Moon Biosystems and used for fluorescent imaging.
Results and Discussion Here, we present a method for amplifying a biorecognition event by allocating photoinitiators onto binding sites on a biochip surface, followed by polymerization of a large number of monomers exclusively at each binding location. Detection of biotin–avidin binding represents a model biorecognition event26 that can readily be extended to the sensitive detection of complementary polynucleotide hybridization for use in the diagnosis of infectious disease. Initiator-functionalized oligonucleotide spots covalently coupled to glass surfaces, initially only detectable with the use of a microarray scanner, became highly visible to the unaided eye following amplification. A schematic of the process using the visible initiation amplification system is presented in Figure 1. To investigate polymerization-based amplification on glass surfaces, the number of propagation reactions generated from each binding event was estimated and reported as an amplification factor. Sensitivity was also measured by characterizing the lowest amount of biotin required to generate a macroscopically detectable response. Results are presented herein. Fluorescent Imaging of Photoinitiators on Surface. As an initial characterization of the amplification processes, aldehyde functionalized microarray surfaces were spotted with rows of
Visual Detection of Labeled Oligonucleotides
Biomacromolecules, Vol. 9, No. 1, 2008 357
Figure 1. Schematic for photopolymerization for signal amplification from a glass biochip surface. Varied amounts of covalently immobilized, labeled oligonucleotide are present on each row of spots on the chip surface. The photoinitiator is stabilized exclusively to positive sites through biotin-streptavidin binding. In cases where the photoinitiator molecule is fluorescent, the allocation of initiator can be verified and quantified through fluorescence scanning. Upon contact with the appropriate monomer formulation (25% PEG(575)DA/225 mM amine co-initiator in H2O) and exposure to light, a hydrogel is grown exclusively from positive spots where the photoinitiator was allocated such that a hydrogel becomes apparent to the unaided eye. This approach enables the elimination of expensive detection equipment to verify the presence of a targeted, labeled oligonucleotide sequence.
Figure 2. Fluorescent image of a microarray surface containing 3′ biotin-labeled capture sequences and unlabeled capture sequences (gray scale: 325-31 000; scanner: 100 PMT%). Spots are ∼570 µm in diameter. The average fluorescent signal of positive spots containing streptavidin-EITC molecules is 26 000, while unlabeled sequences show low fluorescent counts of 400 and an average background of 350 fluorescent counts, close to machine noise (325). Fluorescent intensities show a uniform signal both within individual spots (standard deviation ) 3260) and between duplicate spots (standard deviation ) 3060).
5′-amino, 3′-biotin functionalized capture sequences and rows of 5′-amine capture sequences without the biotin label. Upon the incubation of 1 µg/mL concentrations of streptavidin-EITC, photoinitator molecules should be localized to the labeled capture sequences with the strong affinity of the biotin-streptavidin complex (Kd ) 10-15). Eosin, a weak fluorophore with a fluorescence quantum yield of 0.19,27 has a strong absorbance at 532 nm, an excitation source used by most conventional microarray scanners. Thus, fluorescent scanning was used to quantify the number of biotin-streptavidin binding events on the surface as well as background from nonspecific protein interactions. Figure 2 shows a fluorescent image of a microarray after streptavidin-EITC incubation but prior to amplification. Upon quantification of the fluorescent intensity at the positive spots minus the background absorbance, there are 480 ( 60 eosin fluorophores per square micrometer. Because the number
of fluorescent eosin molecules per streptavidin protein was determined as 2.3, there is an average of 200 biotin-streptavidin binding events per square micrometer at positive sites on the biochip. The variance in signal intensity is
