Article pubs.acs.org/Langmuir
Adapting Fluorescence Resonance Energy Transfer with Quantum Dot Donors for Solid-Phase Hybridization Assays in Microtiter Plate Format Eleonora Petryayeva, W. Russ Algar,† and Ulrich J. Krull* Chemical Sensors Group, Department of Chemical and Physical Sciences, University of Toronto Mississauga, 3359 Mississauga Road North, Mississauga, ON L5L 1C6, Canada S Supporting Information *
ABSTRACT: Methods have been developed for the solid-phase detection of nucleic acids using mixed films of quantum dots (QDs) and oligonucleotide probes in microtiter plates. Polystyrene microwells were functionalized with multidentate imidazole ligands to immobilize QDs. Oligonucleotide hybridization was transduced using QDs as donors in fluorescence resonance energy transfer (FRET). One detection channel paired greenemitting QD donors with Cy3 acceptors and served as an internal standard. A second detection channel paired red-emitting QDs with Alexa Fluor 647 acceptors and served as the primary detection channel. A selective assay for multiple targets was demonstrated using a 96-well plate format, which combined the advantages of two-plex QD-FRET with the highthroughput capability and convenience of microtiter plates. The assay had excellent resistance to the nonspecific adsorption of DNA and discriminated between fully complementary and single base-pair mismatched sequences with a contrast ratio >2. Under optimal conditions for a single color (green QD) assay format, the limit of detection (LOD) was 4 nM, and the dynamic range was from 20 to 300 nM. In a two-color assay, the detection channel (red QD) exhibited linear response between 4 and 100 nM and a LOD of 4 nM.
■
optical fibers using ligand exchange with multidentate thiol surface ligands and then derivatized with oligonucleotide probes. “Sandwich” hybridization between probe, unlabeled target, and fluorescent dye-labeled reporter oligonucleotides yielded efficient FRET and permitted sensitive and selective detection of targets on the basis of dye/QD emission ratios. The assay format permitted multiplexing on a single substrate without spatial registration, was ensemble compatible, and required only a single excitation source. In contrast to solutionphase (i.e., homogeneous) assays, the solid-phase (i.e., heterogeneous) assay format permitted use of washing steps to facilitate measurement of targets that were within complex sample matrices. A potential drawback of the assay format was the need for a custom detection platform built to interrogate QD-modified optical fibers. A more recent report described the development of a microfluidic chip-based assay, and this format also required a specially configured optical detection system.6 Such custom detection platforms may pose a challenge for routine use of these assay formats in other laboratories, despite the inherent benefits in analytical performance. A simple, more widely accessible platform is needed to advance the role of multiplexed, solid-phase QD-FRET assays in bioanalysis. The most common platforms for solid-phase assays of any kind are microtiter plates. The multiwell (e.g., 96, 384) format is well suited for high-throughput screening and analysis.7
INTRODUCTION Semiconductor quantum dots (QDs) are unique nanomaterials with advantageous optical and physical properties.1,2 Broad and strong absorption spectra, narrow and size-tunable photoluminescence (PL), large “effective” Stokes shifts, and high quantum yields make them particularly favorable in bioanalysis. The size of a QD, which is comparable to many biomacromolecules, has surface area that is sufficiently large to serve as a scaffold and offers additional value in bioanalytical applications. Fluorescence resonance energy transfer (FRET) is a powerful transduction mechanism for assays and biosensors, and QDs are ideal FRET donors.3,4 The tunable PL properties of QDs facilitate control over spectral overlap, while their broad UV−visible absorption allows selection of excitation wavelengths where direct excitation of an acceptor dye is minimized.4 The size and surface area of QDs also permit the centrosymmetric assembly of multiple acceptors to increase FRET efficiency and to provide better sensitivity. The combination of these properties has established QDs as superior FRET donors in comparison to molecular dyes, particularly in multiplexed formats.1 To date, QD-FRET constructs have been used in applications such as the detection of drugs, small molecules, nucleic acids, and enzymatic activity.1 We have previously reported solid-phase QD-FRET assays for the detection of nucleic acids in a multiplexed format.5 The simultaneous detection of up to four oligonucleotide sequences was possible using different combinations of FRET and direct excitation of fluorescence.5 Green- and red-emitting thioalkyl acid-coated CdSe/ZnS QDs were immobilized on fused silica © 2013 American Chemical Society
Received: October 29, 2012 Revised: December 18, 2012 Published: January 8, 2013 977
dx.doi.org/10.1021/la304287v | Langmuir 2013, 29, 977−987
Langmuir
Article
Automated microplate readers are standard equipment in many research and clinical laboratories, and the microtiter plate format is widely used for clinical tests (e.g., immunoassays), pharmaceutical research, activity-based protein profiling, cellbased assays, and much more. For these reasons, the integration of nanomaterials and their unique optical properties with the tried-and-true microtiter plate format is an important step toward the practical use of nanotechnology in bioanalytical applications. Recent examples include multiplexed, microtiter plate-based immunoassays that utilize QDs as labels (without FRET),8,9 and ELISA-type gold nanoparticle-assisted detection of pesticides10 or cancer markers (e.g., carcinoembryonic antigen) in human plasma.11 Here, we describe new work where a two-plex solid-phase QD-FRET hybridization assay is adapted to a microtiter plate format for use with a commercial fluorescence plate reader. As shown in Figure 1, microtiter plate wells were derivatized with
the overall capacity of this formata series of parallel two-plex assays across multiple wellsis limited only by the number of wells available (i.e., ≤96, ≤384, or more depending on the plate size; for brevity, we refer to the use of two-plex assays across multiple wells as “multiplexed”). In addition to proof-ofconcept, we address aspects such as optimization of the ratio of immobilized QDs, the effect of the hydrophilic ligand coating on QDs in controlling probe density and hybridization efficiency, passivation to alleviate nonspecific adsorption, single base-pair mismatch discrimination, and surface regeneration. The assay format is very promising for facilitating broad access, by the scientific community at large, to the unique advantages of QDs in bioanalysis.
■
EXPERIMENTAL SECTION
Details about experimental procedures, materials, reagents, and instrumentation can be found in the Supporting Information. Quantum Dots and Oligonucleotides. Hydrophobic CdSe1−xSx/ZnS QDs (Cytodiagnostics, ON, Canada) with emission maxima at 525 nm (gQD) and 624 nm (rQD) were rendered watersoluble by ligand exchange with mercaptopropionic acid (MPA) or glutathione (GSH), as described in the Supporting Information. Probe and complementary target oligonucleotides were obtained from Integrated DNA Technologies (Coralville, IA). Probe oligonucleotides were received in disulfide (DTPA) form and reduced with a 1000-fold excess of TCEP to the dithiol form. The oligonucleotide sequences are listed in Table 1. Modification of Polystyrene Microtiter Plate Wells with Multidentate Ligands. Polystyrene (PS) microtiter plates were oxidized by treatment with KMnO4 in sulfuric acid (0.25 g mL−1) by adding 200 μL/well and incubating for 1 h.14 The solution was removed, and the wells washed five times with 6 M HCl (200 μL/ well), incubating each time for 10 min. Once any insoluble manganese oxide salts were removed, the plates were washed with copious amounts of deionized water, rinsed once with methanol, and dried in air. The newly carboxylated wells were immediately activated with Nhydroxysuccinimide (NHS) ester and subsequently grafted with poly(ethyleneimine) (PEI, MW 2K). First, a solution of 0.1 M N(3-(dimethylamino)propyl)-N′-ethylcarbodiimide (EDC) and 0.2 M NHS in 2-(N-morpholino)ethanesulfonic acid (MES) buffer (pH 5, 100 mM) was dispensed into wells (100 μL/well). The plate was sealed and agitated for 90 min prior to removal of the EDC/NHS solution. Next, the wells were filled with 100 μL/well of aqueous PEI (5 mg mL−1) that had been adjusted to pH 7.5 with 6 M HCl. The plate was then incubated at room temperature for 4−6 h, washed five times with deionized water, once with methanol, and finally air-dried. PEI-modified microtiter wells were further modified with alternating layers of oppositely charged poly(acrylic acid) (PAA; MW ∼ 4500) and poly(ethyleneimine) (PEI; MW ∼ 750K) using the electrostatic layer-by-layer (LBL) technique and subsequent covalent crosslinking.15 Wells were incubated for 30 min with 200 μL/well of PAA solution (1 g L−1) containing 0.5 M NaCl and adjusted to pH 7.4. The PAA solution was discarded, and wells were washed five times with water (200 μL/well) and incubated for 30 min with PEI solution (1 g L−1) containing 0.5 M NaCl at pH 7.4. Wells were washed with water and subjected to two more cycles of PAA and PEI layer deposition (five polymer layers total). For cross-linking between the layers, the wells were incubated with 0.1 M EDC solution (50 μL/ well) for 30 min. Next, the terminal PEI layer was coupled to a polymer with a poly(acrylic acid) backbone and pendant imidazole groups (PAAI). The PAAI was prepared in DMSO as described previously,16 diluted with water (1:1 v/v), and any precipitates were removed by filtration. The PAAI-containing filtrate was activated with EDC (0.1 M), and 100 μL was added to each well. Plates were sealed and incubated for 24 h on an orbital shaker. Wells were rinsed five times with deionized water, once with methanol, and air-dried. The PAAI-modified microtiter plates were stored over desiccant until needed.
Figure 1. Schematic design of nucleic acid hybridization assay developed in microtiter plate format. Water-soluble, thioalkyl acid ligand-capped QDs were immobilized via multidentate imidazole surface ligands coated on polystyrene wells and conjugated with dithiol-modified oligonucleotide probes. Transduction of hybridization was achieved using Cy3 and A647 modified oligonucleotide targets through FRET sensitization of dye fluorescence. The gQD−Cy3 channel was used as internal standard; the rQD−A647 channel was used to assay targets. High-throughput was achieved via the array format of the microtiter plate wells.
multidentate imidazole-based surface ligands that allowed for concurrent immobilization of two colors of QDs via spontaneous metal-affinity coordination in each microtiter well. Immobilized green- and red-emitting QD donors (gQD and rQD) were modified with two different oligonucleotide probe sequences and, through hybridization with target oligonucleotides, were paired with Cy3 and Alexa Fluor 647 acceptor dyes, respectively, to create two FRET-based detection channels. Analogous to previous QD-FRET assays, the ratio of QD donor and FRET-sensitized acceptor dye emission was used for quantitative analysis. This design is complementary to the array format of microtiter plate wells and provides a robust platform for high-throughput analysis by using the gQD−Cy3 channel as an internal standard and the rQD−A647 to assay different target sequences across multiple microtiter wells. We demonstrate that various sample mixtures containing up to three nucleic acid sequences can be assayed concurrently, with an internal standard, and across multiple wells in a 96-well microtiter plate. While multiplexing solely on the basis of discrete spectral channels is limited to ≤8 different analytes,12,13 978
dx.doi.org/10.1021/la304287v | Langmuir 2013, 29, 977−987
Langmuir
Article
Table 1. Oligonucleotidesa name probe target (3′-TRG) target (5′-TRG) 3BPM 1BPM NC−Cy3 probe target probe target probe target NC−A647
sequence Seq 1 DTPA-5′-ATT TTG TCT GAA ACC CTG T-3′ Cy3-3′-TAA AAC AGA CTT TGG GAC A-5′ 3′-TAA AAC AGA CTT TGG GAC A-5 ′-Cy3 Cy3-3′-TGA AAC AGG CTT TGG GAT A-5′ Cy3-3′-TAA AAC ACA CTT TGG GAC A-5′ 5′-ATT TTG TCT GAA ACC CTG T-3′-Cy3 Seq 2 DTPA-5′-CTT ACT TCC ATG ATT TCT TTA ACT-3′ 3′-GAA TGA AGG TAC TAA AGA AAT TGA-5′-A647 Seq 3 DTPA-5′-TTC AGT TAA TCC TAC AAC TTT TTC-3′ 3′-AAG TCA ATT AGG ATG TTG AAA AAG-5′-A647 Seq 4 DTPA-5′-AAC AAT ATT GTC TTG ATT AGT CAT-3′ 3′-TTG TTA TAA CAG AAC TAA TCA GTA-5′-A647 3′-TCA ATT TCT TTA GTA CCT TCA TTC-5′-A647
a
Cy3 - cyanine 3; A647 = Alexa Fluor 647; DTPA = dithiophosphoramidite; NC = noncomplementary; 3BPM = three-base-pair mismatch; 1BPM = one-base-pair mismatch (mismatches are bolded and underlined). All sequences were HPLC purified by the manufacturer. ⎛ ∑λ = d PL(λ) ⎞ ⎛ ∑λ = d PL(λ) ⎞ ⎟ − ⎜ λ=c ⎟ FRET ratio = ⎜⎜ λλ== cb ⎟ ⎜ λ=b ⎟ ⎝ ∑λ = a PL(λ) ⎠DA ⎝ ∑λ = a PL(λ) ⎠D
Immobilization of Quantum Dots. Microtiter wells functionalized with multidentate PAAI ligands were rinsed with 200 μL/well of borate buffer (pH 9.25, 50 mM) and 50 μL/well of aqueous MPA- or GSH-coated QDs (0.05−0.2 μM) were added. The immobilization of two colors of QDs was done from aqueous mixtures of gQD and rQD at a molar ratio of 1.75:1. Wells with immobilized QDs were washed three times with borate buffer, and 50 μL/well of 5 μM probe oligonucleotide solution (borate buffer, pH 8.3, 100 mM; 5−10 mM TCEP) was added. The microtiter plate was sealed and incubated for 4−8 h. Experiments with the immobilization of two probe oligonucleotides were done at a 1:1 ratio, keeping the total probe concentration at 5 μM. After removal of the probe solution, wells were washed three times with borate buffer (pH 9.25) and residual surface active sites were then blocked by adding 50 μL/well of 0.5 mg mL−1 denatured bovine serum albumin (dBSA) in borate buffer (pH 8.3) for 50 min. Wells were then washed with borate buffer (pH 9.25) three times (200 μL/well). The dBSA was prepared using a protocol that was described previously.17 Hybridization Assays. All hybridization assays were done using oligonucleotide target solutions (50 μL) prepared in Tris-borate buffer (TB; pH 7.4, 50 mM, 50 mM NaCl) for 1 h, unless otherwise noted. The target concentrations ranged from 0.2 to 25 pmol/well (4−500 nM). To investigate the discrimination of single base-pair mismatches and the regeneration of QD-probe conjugates, the buffers contained formamide at ratios between 0−25% and 50% (v/v), respectively. Multiplexed assays used 4 pmol/well of Seq 1 (3′-TRG) and the gQD−Cy3 FRET channel as an internal standard; Seq 2, Seq 3, and Seq 4 oligonucleotide targets were used at concentrations as described in the Results and Discussion (see Figure 6b). Target oligonucleotide sequences are listed in Table 1. Seq 1 is diagnostic of the H. sapiens survival motor neuron protein coding gene (SMN 1) and associated with spinal muscular atrophy;18 Seq 2 is associated with E. coli βglucoronidase enzyme coding gene (uidA);18 Seq 3 is associated with a portion of E. coli hemolysin coding gene (hly A);19 and Seq 4 is associated with a portion of the L. monocytogenes listeriolysin O coding gene (hly A).19 Table 1 also lists noncomplementary (NC) sequences and sequences with single (1BPM) and three (3BPM) base pair mismatches. Data Analysis. PL spectra from microtiter plates were collected using a Tecan M1000 microplate reader (Tecan, Inc., Durham, NC) with 405 nm excitation. All PL spectra were background corrected and normalized to the gQD PL emission maxima. A FRET ratio, eq 1, was used for quantitative analysis
(1)
where the wavelength range in the numerators corresponded to a band of acceptor PL and the wavelength range in the denominator corresponded to a band of donor PL; DA denotes in the presence of acceptor and D in its absence. For the gQD−Cy3 FRET pair, the PL bands were defined by a = 510 nm, b = 534 nm, c = 560 nm, and d = 590 nm; for the rQD−A647 FRET pair, the PL bands were defined by a = 615 nm, b = 645 nm, c = 655 nm, and d = 695 nm. The analysis of complex mixtures was done using a standardized FRET ratio (SFR), eq 2, to correct for the response in the internal standard channel: SFR =
■
FRET ratio(rQD−A647) FRET ratio(gQD−Cy3)
(2)
RESULTS AND DISCUSSION Modification of Polystyrene Microtiter Plate Wells with QDs. Reproducible immobilization of QDs within PS wells at high density was crucial to the solid-phase microtiter plate-based QD-FRET assay reported here. Therefore, PS wells were functionalized with high affinity multidentate surface ligands that were covalently bound to the PS surface. This chemistry provided multiple points of interaction with the inorganic shell of the QDs, tightly anchoring them to the surface.20 Since the native PS surface lacked the functional groups needed for covalent modification, it was first subjected to wet chemical treatment with a strong oxidizing agent (potassium permanganate/H2SO4). This generated carboxylic acid groups for attachment of the surface ligands, which incorporated pendant imidazole groups. Over the past several years, the bioconjugation of QDs with imidazole-bearing, hexahistidine-modified biomolecules has become a proven and versatile strategy. Rapid self-assembly of proteins, peptides, and oligonucleotides has been demonstrated by a number of groups with excellent control over the conjugate valence.21 The Bawendi group described a polymer coating with appended imidazole ligands (50%) that facilitated ligand exchange with organic QDs within 10 min at room temperature.22 Recently, we reported QD immobilization on glass substrates using 979
dx.doi.org/10.1021/la304287v | Langmuir 2013, 29, 977−987
Langmuir
Article
multidentate imidazole surface ligands.16 QDs capped with thiol-based ligands were immobilized at high density on fused silica fibers or glass beads and retained their functionality in oligonucleotide hybridization assays and protease assays.16 The glass substrates were amine functionalized and then directly coupled with an imidazole-appended polymer (PAAI) based on a poly(acrylic acid) backbone. Building on the success of this immobilization approach, we have now extended it to the derivatization of PS wells. Oxidized PS wells were initially grafted with PEI and further modified with a cross-linked layerby-layer assembly to achieve a uniformly high density of primary amine functional groups within the wells for covalent coupling with PAAI. Modification of the PS wells was monitored with high resolution angularly resolved X-ray photoelectron spectroscopy (XPS) analysis of signals from C 1s, N 1s, and O 1s to confirm incorporation of functional groups on the surface (see Supporting Information for details). PS wells functionalized with multidentate PAAI ligands were used to immobilize gQDs coated with either MPA or GSH ligands. QD immobilization was typically done over 4−6 h and helped ensure homogeneous coverage and reproducibility between wells. The variation in QD PL intensity was
