Paper-Based Solid-Phase Multiplexed Nucleic Acid Hybridization

Jul 9, 2013 - ... of Toronto Mississauga, 3359 Mississauga Road North, Mississauga ON, L5L 1C6, Canada .... Analytical Chemistry 2016 88 (3), 1768-177...
0 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/ac

Paper-Based Solid-Phase Multiplexed Nucleic Acid Hybridization Assay with Tunable Dynamic Range Using Immobilized Quantum Dots As Donors in Fluorescence Resonance Energy Transfer M. Omair Noor 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: A multiplexed solid-phase nucleic acid hybridization assay on a paper-based platform is presented using multicolor immobilized quantum dots (QDs) as donors in fluorescence resonance energy transfer (FRET). The surface of paper was modified with imidazole groups to immobilize two types of QDprobe oligonucleotide conjugates that were assembled in solution. Green-emitting QDs (gQDs) and red-emitting QDs (rQDs) served as donors with Cy3 and Alexa Fluor 647 (A647) acceptors. The gQD/Cy3 FRET pair served as an internal standard, while the rQD/A647 FRET pair served as a detection channel, combining the control and analytical test zones in one physical location. Hybridization of dye-labeled oligonucleotide targets provided the proximity for FRET sensitized emission from the acceptor dyes, which served as an analytical signal. Hybridization assays in the multicolor format provided a limit of detection of 90 fmol and an upper limit of dynamic range of 3.5 pmol. The use of an array of detection zones was designed to provide improved analytical figures of merit compared to that which could be achieved on one type of array design in terms of relative concentration of multicolor QDs. The hybridization assays showed excellent resistance to nonspecific adsorption of oligonucleotides. Selectivity of the two-plex hybridization assay was demonstrated by single nucleotide polymorphism (SNP) detection at a contrast ratio of 50:1. Additionally, it is shown that the use of preformed QD-probe oligonucleotide conjugates and consideration of the relative number density of the two types of QD-probe conjugates in the two-color assay format is advantageous to maximize assay sensitivity and the upper limit of dynamic range. diagnosis and screening.10 Practical advantages include fluid transport by capillary action, a capability of filtering of samples to remove aggregates and particles, surface chemistry that can readily be modified, good compatibility with biological samples, and ease of safe disposal by incineration to eradicate biohazards.11,12 For solid-phase assays that make use of QDs on PADs, the residual ash that is generated upon incineration can be collected and handled as heavy metal waste. From the standpoint of solid-phase nucleic acid hybridization assays using arrays, it has been reported that variations of spotto-spot probe density and heterogeneity within a spot can influence the efficiency and kinetics of target hybridization.13,14 This suggests that standardization that is achieved by reliance on a control zone that is physically separated from the test zone in PADs is unable to account quantitatively for signal variations associated with heterogeneity in interfacial chemistry. The microzone device format for PADs is based on spatially resolved zones and is useful for high-throughtput screening and multiplexed analysis.15,16 However, no work has appeared in the

T

he unique optical properties of colloidal semiconductor nanocrystals or quantum dots (QDs) that are useful for optical multiplexing include broad absorption spectra; narrow, symmetric, and size-tunable photoluminescence (PL) spectra that offer large Stokes shifts; long PL lifetimes (>20 ns); and good photostability.1 It is the combination of such properties that serves to make QDs excellent donors in fluorescence resonance energy transfer (FRET).2 In a FRET based transduction scheme, QDs can be used as integrated nanomaterials for the development of FRET-based biosensors and bioassays, where signal development from selective interactions of probes on the surface of QDs results in modulation of QD PL.3 A physical configuration that allows multiple acceptors to interact with a single QD donor provides improved FRET efficiency and analytical sensitivity.4 Further advantages include the potential of implementation of a ratiometric approach for detection and analysis to improve precision5,6 and that solidphase QD-FRET assays offer higher sensitivity,7 reusability,8,9 and physical containment of QDs. The use of paper as a solid support for assay development has received renewed attention in recent years. Paper-based analytical devices (PADs) ideally offer low cost and ease of use and hence are attractive for their applicability for point-of-care © XXXX American Chemical Society

Received: May 17, 2013 Accepted: July 9, 2013

A

dx.doi.org/10.1021/ac401471n | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

Figure 1. (a) Synthetic steps involved in the surface modification of paper (cellulose) with the imidazole ligands for the immobilization of greenemitting QDs (gQDs) and red-emitting QDs (rQDs) that were modified with SMN1 oligonucleotide probes and uidA oligonucleotide probes respectively. (b) Design of the paper-based solid-phase multiplexed nucleic acid hybridization assay using multicolor immobilized QDs as donors in FRET. Hybridization with Cy3 labeled SMN1 and A647 labeled uidA target oligonucleotides provided the proximity for FRET sensitized emission from Cy3 and A647 dyes upon excitation with a 402 nm laser, which served as an analytical signal.

■

EXPERIMENTAL SECTION A detailed description of reagents, experimental procedures, and instrumentation used in these experiments can be found in the Supporting Information. The oligonucleotide sequences used in the hybridization assays are listed in Table 1.

literature describing the integration of internal standardization within each zone of selective chemistry for target analyte detection, i.e., merging of the control zone and test zone within one physical location. The use of QDs as donors in FRET for a ratiometric detection of acceptor dye-labeled oligonucleotide targets on a paper-based platform was recently reported by our group8 and offered a significant advantage in amelioration of nonspecific adsorption. In this new work, we explore a two-plex solid-phase QD-FRET nucleic acid hybridization assay on a paper-based platform and demonstrate that one of the FRET color channels can serve as a control while the second FRET color channel can provide for an analytical signal. The schematic of the assay design is depicted in Figure 1. The surface of paper was modified with imidazole groups (Figure 1a) to immobilize two colors of QD-probe oligonucleotide conjugates. The QD-probe conjugates were preassembled in solution. The green-emitting (gQDs) and red-emitting QDs (rQDs) were modified with SMN1 and uidA oligonucleotide probes, respectively. Immobilized gQDs and rQDs served as FRET donors with Cy3 and Alexa Fluor 647 (A647) acceptor dyes, respectively, to create two FRET detection channels. Addition and subsequent hybridization of dye-labeled oligonucleotide targets provided the proximity for FRET sensitized dye emission upon excitation of QD-probe conjugates, which served as an analytical signal (Figure 1b). A ratiometric detection approach based on QD donor PL quenching and sensitization of acceptor-dye emission upon target hybridization was used for quantitative analysis. The gQD/Cy3 (donor/acceptor) FRET channel was used as an internal standard, and the rQD/A647 FRET channel was used as the target detection channel. This provided for internal standardization within each zone that was placed on the paper substrates. The hybridization assays showed excellent resistance to nonspecific adsorption of oligonucleotides and discriminated between a fully complementary and one base pair mismatched sequences with a contrast ratio of greater than 50.

Table 1. Oligonucleotide Sequences Used in the Hybridization Assaysa gQD/Cy3 FRET pair (SMN1 sequences) SMN1 SMN1 SMN1 SMN1

probe DTPA - 5′- ATT TTG TCT GAA ACC CTG T - 3′ FC TGT Cy3 - 3′- TAA AAC AGA CTT TGG GAC A - 5′ 1 BPM TGT Cy3 - 3′- TAA AAC ACA CTT TGG GAC A - 5′ NC TGT Cy3 - 3′- TGT CCC AAA GTC TGT TTT A - 5′ rQD/A647 FRET pair (uidA sequences)

uidA probe uidA FC TGT uidA NC TGT

DTPA - 5′- CTT ACT TCC ATG ATT TCT TTA ACT 3′ A647 - 3′- GAA TGA AGG TAC TAA AGA AAT TGA - 5′ 3′- TTG TTA TAA CAG AAC TAA TCA GTA - A647 - 5′

a

TGT = target; FC = fully complementary; 1 BPM = 1 base pair mismatch; NC = non-complementary; Cy3 = Cyanine 3; A647 = Alexa Fluor 647; DTPA = dithiol phosphoramidite. The mismatched base in SMN1 1 BPM TGT is bolded and underlined.

Preparation of QD-Probe Oligonucleotide Conjugates. Oleic acid capped CdSxSe1−x/ZnS (core/shell) QDs in toluene were made water-soluble by a ligand exchange reaction with glutathione (GSH).8 GSH capped gQDs (GSHgQDs, peak PL at 525 nm) and GSH capped rQDs (GSHrQDs, peak PL at 614 nm) were subsequently modified with SMN1 oligonucleotide probes (SMN1 probe) and uidA oligonucleotide probes (uidA probe), respectively (see Supporting Information for the experimental details). The “mixed-films” of oligonucleotide probes on QDs consisted of both colors of QDs (gQDs and rQDs) in a mixture that was B

dx.doi.org/10.1021/ac401471n | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

oligonucleotide probes (QD-probe conjugates) as one construct on the surface of a paper-based solid support that is functionalized with imidazole ligands.8 This latter strategy was used in this work for the assembly of the biorecognition interface that consisted of immobilized QD-probe conjugates. The use of imidazole ligands for the immobilization of QDs or QD-probe conjugates was also reported to be advantageous from the standpoint of retention of solution-phase optical properties of QDs, such as quantum yield, peak PL, and fullwidth-at-half-maximum of QD PL.8,23 The oligonucleotide probes were terminated with disulfide (DTPA) at the 5′ terminus, and this was reduced to a dithiol form using tris(2-carboxyethyl)phosphine hydrochloride. This was followed by in situ conjugation to the surface of the ZnS shell of GSH-QDs by self-assembly via a ligand exchange reaction, as reported previously.8,9,21,23 The conjugation of SMN1 probes and uidA probes to the surface of gQDs and rQDs, respectively, was confirmed by gel electrophoresis and solution-phase hybridization experiments (see Figures S2 and S3 in the Supporting Information). Single-Color Hybridization Assays. Paper-based solidphase assays are advantageous for rapid transduction of nucleic acid hybridization. We have previously reported that the signal profiles associated with target hybridization can reach a steady state in less than 2 min for a solid-phase QD-FRET nucleic acid hybridization assay.8 The capillary wicking action in a paperbased matrix serves as an active delivery method to move target oligonucleotides to the biorecognition elements, which facilitates fast hybridization kinetics by overcoming diffusionlimited kinetics.24 Initial experiments were conducted in a single-color format (one FRET pair) with each of the two FRET pairs to provide for quantitative transduction of nucleic acid hybridization. The SMN1 sequence is a diagnostic of neuromuscular disease spinal muscular atrophy, while the uidA sequence is a diagnostic of Escherichia coli.18 A FRET ratio was used to quantify the signal that was associated with the amount of target DNA added (see eqs S4 and S5 in the Supporting Information). FRET ratio is a ratiometric approach for signal quantification and takes into account donor PL quenching and sensitization of acceptor PL. This approach for quantification is advantageous for analytical measurements as it is less susceptible to variations imposed by sample preparation and changes in the instrumental response.6,25,26 Figure 2a shows a FRET ratio plot for the response of the assay to increasing amounts of SMN1 FC TGT with the gQD/ Cy3 FRET pair. For the corresponding PL images, see Figure S4. A linear increase in the FRET ratio response of the assay for target amounts ranging from 1.5 pmol to 12 pmol is observed, which corresponds to a dynamic range of ca. 1 order of magnitude. A dynamic range of 1 order of magnitude has previously been reported for solid-phase QD-FRET nucleic acid hybridization assays done using glass beads,9 microfluidics,7 optical fibers,18−20,27 and paper.8 The limit of detection (LOD) of the assay was experimentally determined to be ca. 150 fmol (50 nM) and corresponded to a FRET ratio response that was 3 standard deviations higher than the average background FRET ratio signal. Figure 2b shows a FRET ratio plot for the response of the assay to increasing amounts of uidA FC TGT with the rQD/A647 FRET pair. The corresponding PL images are shown in Figure S4. A linear increase in the FRET ratio response of the assay for target amounts ranging from 0.75 pmol to 9.0 pmol is observed, which corresponds to a

concurrently modified with the two probe types (SMN1 and uidA). Surface Modification of Paper with Imidazole Groups. Whatman cellulose chromatography paper (Grade 1) cut to dimensions of 25 mm by 60 mm was modified with an aldehyde functionality using periodate oxidation of cellulose as reported elsewhere.17 Aldehyde functionalized paper was subsequently modified with imidazole groups in an array format (array size was 3 × 6; spot diameter was 5 mm) using a previously published protocol.8 Details can be found in the Supporting Information. Immobilization of QD-Probe Oligonucleotide Conjugates. Prior to the immobilization of QD-probe conjugates on imidazole modified paper, the surface of paper was blocked with 0.1% bovine serum albumin (BSA) in 100 mM tris-borate buffer (TB, pH 7.44) for 30 min. The paper was then washed with 50 mM borate buffer (BB, pH 9.25) for 10 min. For single-color experiments (one donor/acceptor FRET pair), each element of an imidazole modified array was subjected to manual pipetting with a 2 μL aliquot of a solution of either gQD-SMN1 probe conjugates or rQD-uidA probe conjugates at 400 nM concentration. For the two-color experiments (both FRET pairs), solutions of gQD-SMN1 probe conjugates and rQD-uidA probe conjugates at ca. 400 nM each were mixed in equal proportions prior to their application on the imidazole modified elements of an array. For the two-color experiments in the presence of mixed-films, a 2 μL aliquot of a mixed-film solution was spotted on each element of imidazole modified array of the paper. The spotted papers were stored in the dark at room temperature for 2 h. The paper was then washed with BB buffer for 10 min. Hybridization Assays and Selectivity. Hybridization experiments were done by pipetting a 3 μL aliquot of a solution of oligonucleotide targets at various concentrations (16 nM to 6 μM) onto the elements (zones) where QD-probe conjugates were immobilized. The hybridization experiments were done in 50 mM BB buffer saline (BBS, pH 9.25, 100 mM NaCl), and the target solution was allowed to sit for 30 min at room temperature prior to image scans and collection of PL spectra without any washing of the paper. For SNP discrimination experiments, the paper was initially exposed to FC and 1 BPM targets in BBS buffer and subsequently exposed to 10% (v/v) formamide in BB buffer for 10 min.

■

RESULTS AND DISCUSSION The FRET Pairs. The two FRET pairs used in this work consisted of gQD/Cy3 (donor/acceptor) and rQD/A647 (donor/acceptor). The Förster formalism was used for insolution characterization of the two FRET pairs (see Supporting Information). The absorption and PL spectra of the two FRET pairs are shown in the Supporting Information (Figure S1). Assembly of the Biorecognition Interface. A typical approach for the assembly of a biorecognition interface for the development of solid-phase QD-FRET nucleic acid hybridization assays has been the immobilization of QDs on the surface of a solid support followed by conjugation of oligonucleotide probes to the surface of immobilized QDs. This approach has been used for the development of solidphase QD-FRET nucleic acid hybridization assays on solid supports such as optical fibers,18−20 glass beads,9 microtiter plates,21 and microfluidics.7,22 We have recently reported that it is possible to immobilize QDs that are modified with C

dx.doi.org/10.1021/ac401471n | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

complementary (NC) oligonucleotide targets. The results for the exposure of NC oligonucleotide targets to the gQD/Cy3 and rQD/A647 FRET pairs are shown in Figure 3a. For the corresponding PL spectra and images, see Figure S7 in the Supporting Information.

Figure 2. Single-color quantitative hybridization assays. Response of the (a) gQD/Cy3 and (b) rQD/A647 FRET pairs to increasing amounts of SMN1 FC TGT and uidA FC TGT, respectively. Insets show PL spectra corresponding to their respective data points in a and b.

Figure 3. Selectivity of solid-phase nucleic acid hybridization assays demonstrated with the (a) gQD/Cy3 and rQD/A647 FRET pairs and (b) SNP discrimination with the gQD/Cy3 FRET pair. (a) FRET ratios for the hybridization of SMN1 FC, SMN1 NC, uidA FC, and uidA NC targets. The black arrows in a indicate the FRET ratio signal for the hybridization of NC targets. (b) FRET ratios for the hybridization of SMN1 FC and SMN1 1 BPM targets in BBS buffer and after exposure to BB buffer containing 10% (v/v) formamide (F). The black arrow indicates SNP discrimination.

dynamic range of ca. 1 order of magnitude. The LOD of the assay in the case of the rQD/A647 FRET pair was ca. 240 fmol (80 nM). It is interesting to note that the two FRET pairs yielded two different values of the upper limit of the dynamic range (limit of linearity, LOL), 12 pmol in case of the gQD/Cy3 FRET pair, and 9.0 pmol in case of the rQD/A647 FRET pair. Absorbance measurements showed that on average eight to nine SMN1 probes and six to seven uidA probes were conjugated to the surface of GSH-gQDs and GSH-rQDs, respectively (see Figure S5). Solution-phase hybridization experiments showed that the FRET ratio response of the gQD/Cy3 FRET pair reached saturation at an acceptor to QD ratio of 6:1, while the FRET ratio response of the rQD/A647 FRET pair reached saturation at an acceptor to QD ratio of 5:1 (see Figure S6). These results suggest that the LOL of the single-color assays for the two FRET pairs is governed by the saturation of the FRET efficiency response instead of the saturation of the biorecognition element, as reported elsewhere.9,28 It is important to note that the dimensions of the gQDs and rQDs (diameter of 6−7.5 nm) used in this study are the same. The different emission properties of the ternary alloy QDs result from changing the composition of the core.29 Additionally, during the preparation of QD-probe conjugates, both colors of QDs were incubated with a 20-fold molar excess of oligonucleotide probes. Nonspecific Adsorption and SNP Detection. The extent of nonspecific adsorption of oligonucleotides on the surface of immobilized QD-probe conjugates was evaluated using non-

We have previously reported that the implementation of GSH-QDs for the development of solid-phase QD-FRET nucleic acid hybridization assay provides excellent resistance to nonspecific adsorption of oligonucleotides.8 The two primary factors responsible for providing excellent resistance to nonspecific adsorption of oligonucleotides are the zwitterionic character of GSH23 and the pH of the buffer (pH 9.25) that was used for the hybridization assays.30,31 It is important to note that no washing of the paper was done after an application of oligonucleotide target solution to the zone with immobilized QD-probe conjugates. In the case of the gQD/Cy3 FRET pair, the exposure of 12 pmol of SMN1 FC TGT yielded a FRET ratio response of 1.56 (±0.13), while the same amount of SMN1 NC TGT yielded a FRET ratio response of 7.70 (±8.33) × 10−4. This corresponds to a contrast ratio of 2031:1 for the FC and NC targets, respectively. In the case of the rQD/A647 FRET pair, the exposure of 12 pmol of uidA FC TGT yielded a FRET ratio response of 0.40 (±0.02), while the same amount of uidA NC TGT yielded a FRET ratio response of 9.3(±2.4) × 10−3. This corresponds to a contrast ratio of 43:1 for the FC and NC targets, respectively. D

dx.doi.org/10.1021/ac401471n | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

(ca. 200 nM) for the assembly of the biorecognition interface for the two-color hybridization assays. The two-color immobilization of gQD-SMN1 probe conjugates and rQDuidA probe conjugates is presented in Figure S9 in the Supporting Information. Singleplex Hybridization Assays in the Two-Color (Multicolor) Format. A significant limitation of the previous solid-phase QD-FRET multiplexed nucleic acid hybridization assays has been the lack of control of the conjugation of one type of oligonucleotide probe to a particular color of QD. This limitation arose from the two-step sequential assembly of the transduction interface that has typically been used, where multiple colors of QDs were first immobilized on the surface of a solid support, followed by indiscriminate conjugation of different types of oligonucleotide probes to the surface of immobilized QDs.18,20−22 The aforementioned sequential assembly of the transduction interface resulted in a mixedfilm of different types of oligonucleotide probes conjugated to the surface of one color of QD. As a result, upon introduction of acceptor-labeled oligonucleotide targets, some of the acceptors were located on the surface of QDs that could not serve as donors, and this concomitantly reduced the sensitivity of each detection channel, i.e., FRET ratio response of each detection channel. This is linked to donor (one color of QDs vs multicolor QDs immobilized as a monolayer) and probe (acceptor) dilutions (one type of probe sequence vs different types of probe sequences immobilized at the transduction interface). The donor and probe dilution effects reduced the effective densities of one color of QD and one type of oligonucleotide probe sequence at the transduction interface. This resulted in reduction of energy transfer pathways with a concomitant decrease in the FRET ratio response for each detection channel. The effects of donor and acceptor dilutions are inherent and unavoidable in the multiplexed assay configuration. The presence of mixed films of different types of probes on the same color of QD can be avoided and has not been previously reported. Herein, the limitation of mixed-films of different types of probes on the same color of QD was overcome by preassembly of the QD-probe oligonucleotide conjugates, which were then immobilized onto the surface of imidazolemodified paper. The preassembly of conjugates in solution allowed for the control of the conjugation of only one type of oligonucleotide probe to a particular color of QD (gQDs modified with SMN1 probes and rQDs modified with uidA probes) without the need to resort to orthogonal chemistries to achieve control of conjugation. In order to gain insight as to whether the conjugation of one type of oligonucleotide probe to a particular color of QD provides any advantage in terms of the analytical figures of merit, singleplex hybridization experiments (one type of oligonucleotide target, SMN1 or uidA) were conducted in the multicolor assay format (both types of QD-probe conjugates immobilized on the surface of paper) in the absence and presence of mixed-films of oligonucleotide probes. In the absence of mixed-films of oligonucleotide probes, gQDs were modified with SMN1 probes only and rQDs were modified with uidA probes only, and then these conjugates were immobilized on the surface of imidazole-modified paper. For mixed-films of oligonucleotide probes, both colors of QDs were modified with both the SMN1 and uidA oligonucleotide probes (see Figure S2), and then the conjugated QDs were immobilized. The results for the hybridization assays are

Solution-phase hybridization assays done with FC and NC oligonucleotide targets provided the selectivity contrast ratios of 30:1 and 5:1 for the gQD/Cy3 and the rQD/A647 FRET pairs, respectively (see Figure S3 in the Supporting Information). This corresponds to ca. 68 fold and ca. 9 fold higher contrast ratios for the solid-phase assays as compared to the solution-phase assays for the gQD/Cy3 and rQD/A647 FRET pairs, respectively. The higher contrast ratios observed with the solid-phase assays can be attributed to the enhancement of FRET efficiency that has been previously reported for the solid-phase QD-FRET nucleic acid hybridization assay.7,8 Overall, these results demonstrate that solid-phase QD-FRET assays are advantageous from the standpoint of improved analytical sensitivity. Many nucleic acid diagnostic applications, such as determination of the presence and copy number of genetic mutations, require mismatch discrimination at single base pair level. In the work herein, a combination of ionic strength and formamide concentration was used to tune the stringency of nucleic acid hybridization for single nucleotide polymorphism (SNP) discrimination. Lowering the ionic strength destabilizes a DNA duplex by preventing the charge screening that is required to stabilize the two strands of a hybrid,32,33 while formamide (F) lowers the melt temperature of a DNA duplex by serving as a hydrogen bond disrupter.34 The use of formamide and ionic strength to control the stringency of nucleic acid hybridization for SNP discrimination is advantageous as compared to the temperature control due to the temperature dependence of the quantum yields of QDs and molecular fluorophores.35 Additionally, this method allows SNP discrimination to be achieved under room temperature conditions without the need for external heaters. Hybridization assays for SNP discrimination were conducted with gQD/Cy3 FRET pair, and the results are shown in Figure 3b. In the absence of any stringency control, i.e. hybridization assays conducted in BBS buffer, a negligible contrast ratio of 1.24:1 is seen between SMN1 FC TGT and SMN1 1 BPM TGT. Incubation of the paper for 10 min in BB buffer containing 10% (v/v) F provided a contrast ratio of ca. 70:1 for SMN1 FC and SMN1 1 BPM targets. For the corresponding PL spectra and images, see Figure S8. These results are in agreement with our previously published study where a 10 min incubation of the paper in BB containing 10% (v/v) F provided an excellent contrast ratio for SNP discrimination for the same SMN1 probe and target sequences used in this work.8 Two-Color Immobilization. Optical multiplexing for a solid-phase QD-FRET nucleic acid hybridization assay is commonly done by implementing a distinct QD/molecular fluorophore (donor/acceptor) FRET pair for each target sequence, which requires coimmobilization of multicolor QDs and oligonucleotide probe sequences. A criterion that is commonly used for the optimization of coimmobilization of multicolor QDs is to match their PL intensities.18,20 This is done to account for the differences in the absorption cross sections and quantum yields of multicolor QDs.36 The alloy QDs used in this work have similar quantum yields (QYs) and absorption coefficients (ε) at the excitation wavelength (ε of GSH-gQDs and GSH-rQDs are 1.1 × 106 M−1 cm−1 and 1.2 × 106 M−1 cm−1 at 402 nm, respectively; QYs of GSH-gQDs and GSH-rQDs are 67% and 77%, respectively). As a result, in order to obtain comparable PL intensities from the two colors of QDs, gQD-SMN1 and rQDuidA probe conjugates were mixed in equimolar concentrations E

dx.doi.org/10.1021/ac401471n | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

(±0.021) pmol−1. This corresponds to ca. 2 fold higher assay sensitivity in the multicolor format as compared to the singlecolor format. A similar trend was observed with the rQD/A647 FRET pair, where single-color (Figure 2b) and multicolor assay formats (Figure 4b) provided assay sensitivities of 0.056 (±0.003) pmol−1 and 0.132 (±0.004) pmol−1, respectively. The multicolor assay format (absence of mixed-films of probes) provided ca. 2.3 fold higher assay sensitivity as compared to the single-color assay format. These results can be explained by the fact that in ratiometric QD-FRET based assays, the FRET sensitized acceptor PL (the analytical signal) exists in the bright background of QD donor PL.9 In the multicolor assay format, this background is reduced as compared to the single-color assay format due to QD dilution. The number of particular colors of QDs interrogated in the multicolor assay format is reduced as compared to the single-color assay format. As a result, the QD PL background signal in the multicolor assay format is lower than the singlecolor assay format, and this provides improved assay sensitivity (higher FRET ratio) in the multicolor assay format. It is also interesting to note that the dilution of each color of QD by a factor of 2 (1:1 dilution for both colors of QDs) provided almost 2-fold improvement in the assay sensitivity in the multicolor assay format as compared to the single-color assay format. In terms of comparing the upper limit of dynamic range (limit of linearity, LOL), the multicolor assay format is expected to provide lower LOL as compared to the singlecolor assay format. This is due to the fact that the number of biorecognition sites (oligonucleotide probes) of a particular sequence type is reduced in the multicolor assay format as compared to the single-color assay format. The experimental results shown in Table S1 are consistent with this hypothesis. In the case of the gQD/Cy3 FRET pair, the LOLs of singlecolor (Figure 2a) and multicolor assay formats (Figure 4a) were 12 pmol and 4.5 pmol, respectively, while in the case of rQD/Cy3 FRET pair, the LOLs of single-color (Figure 2b) and multicolor assay formats (Figure 4b, absence of mixed-film) were 9.0 pmol and 3.0 pmol, respectively. It is interesting to note that the LOLs in the multicolor assay format were reduced by greater than 2 fold (almost 3 fold) as compared to the single-color assay format, despite 50% dilution of each color of QD-probe conjugates. These results again highlight that the LOL in the QD-FRET assay is dependent not only on the saturation of the biorecognition element but also on the saturation of the FRET efficiency response. The results for singleplex hybridization assay in multicolor format in the presence of a mixed-film of oligonucleotide probes with the rQD/A647 FRET pair (uidA FC TGT) are shown in Figure 4c. As compared to Figure 4b (absence of mixed-films), it can be seen that the presence of a mixed-film of oligonucleotide probes not only reduces the sensitivity of the assay but also lowers the dynamic range of the assay. The axes scales in Figure 4b and c have been kept the same for comparison purposes. The assay sensitivities in the presence and absence of a mixed film of oligonucleotide probes were 0.088 (±0.015) pmol −1 and 0.132 (±0.004) pmol −1 , respectively. The control of the conjugation of one type of oligonucleotide probe to one color of QD (absence of mixedfilms) provided 1.5 fold higher assay sensitivity as compared to the mixed-films of oligonucleotide probes. This is expected as in the presence of a mixed film of oligonucleotide probes, some of the acceptors (A647 labeled uidA FC TGT) were located on

shown in Figure 4, and the analytical figures of merit for the single-color hybridization assays and the singleplex hybrid-

Figure 4. Quantitative single target (SMN1 or uidA) hybridization assays in the two-color format with the (a) gQD/Cy3 and (b) rQD/ A647 FRET pairs without mixed-films and with the (c) rQD/A647 FRET pair with mixed-films of oligonucleotide probes. Note: the axes scales in b and c are kept the same for comparison purposes. Insets show PL spectra corresponding to the data points.

ization assays in the multicolor assay format in the presence and absence of mixed-films of oligonucleotide probes are summarized in Table S1. Comparing the response of the gQD/Cy3 FRET pair (SMN1 FC TGT) in the single-color (Figure 2a) and multicolor assay formats (Figure 4a), it is interesting to note that the sensitivity of the assay in the multicolor format is higher than the single-color format, despite donor and probe dilutions in the multicolor case. In the case of the single-color assay format, the assay sensitivity was 0.151 (±0.008) pmol−1, while the sensitivity of multicolor assay format was 0.310 F

dx.doi.org/10.1021/ac401471n | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

Figure 5. Tuning the dynamic range response of the rQD/A647 FRET channel by adjusting the relative molar ratio of the gQD-SMN1 probe and rQD-uidA probe conjugates in the elements of array on PAD. Relative molar ratios of the gQD-SMN1 probe and rQD-uidA probe conjugates at (a) 1:9, (b) 1:1, and (c) 9:1. (i) Normalized PL spectra of the immobilized two types of QD-probe conjugates at various aforementioned relative molar ratios. The inset shows a schematic drawing of the relative molar ratio of the two types of QD-probe conjugates. (ii) Quantitative response of the rQD/A647 FRET channel with an increasing amount of uidA FC TGT. (Inset) Response of the assay at fmol quantities of uidA FC TGT.

The results summarized in Table S1 (Figures 2 and 4) suggest the possibility of tuning the dynamic range of a FRET channel by adjusting the relative ratio of the two types of QDprobe conjugates that are immobilized within zones on a PAD. To test this hypothesis, an array of zones with different relative molar ratios of gQD-SMN1 probe conjugates and rQD-uidA probe conjugates were constructed, and the response of the rQD/A647 FRET pair was evaluated with uidA FC TGT. The three relative molar ratios investigated were 1:9, 1:1, and 9:1 for the gQD-SMN1 probe conjugates and rQD-uidA probe conjugates, respectively. The results are shown in Figure 5, and the analytical figures of merit for each of the mole ratios that were examined are summarized in Table 2. As the relative contribution of rQD-uidA probe conjugates decreased in terms of molar ratio, the sensitivity and the LOD of the rQD/A647 channel increased (e.g., compare 1:9 case with the 9:1 case). However, a decrease in relative contribution of the rQD-uidA conjugates was commensurate with a decrease in the LOL. These results are in agreement with the previous results where

the surface of gQDs that could not serve as donors. In terms of the upper limit of dynamic range, the presence and absence of mixed films of oligonucleotide probes provided LOLs of 2.2 pmol and 3.0 pmol, respectively, a 1.4 fold higher LOL in the absence of mixed films of oligonucleotide probes as compared to the presence of mixed films of oligonucleotide probes. It should be noted that the number of each probe type, i.e., the number of SMN1 and uidA probes, in the presence and absence of a mixed film of oligonucleotide probes is the same. However, in the presence of mixed films, SMN1 and uidA probes are equally distributed among gQDs and rQDs, while in the absence of mixed-films, SMN1 probes are only localized on gQDs, and uidA probes are only localized on rQDs. Different LOLs associated with the presence and absence of mixed films of oligonucleotide probes further reinforces that the upper limit of dynamic range in the QD-FRET based assay is not only governed by the saturation of the biorecognition element but also by the threshold number of acceptors that result in saturation of the FRET efficiency response. G

dx.doi.org/10.1021/ac401471n | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

standardization as compared to a spatially segregated approach that has been more typically reported with paper-based assays. From the standpoint of solid-phase nucleic acid hybridization, variations in terms of interfacial chemistry between spatially resolved zones is known to influence the efficiency and kinetics of nucleic acid hybridization, which can affect signal generation.13,14 As a result, a spatially segregated approach is unable to quantitatively account for heterogeneity associated with different zones or pads of interfacial chemistry in a reliable manner. However, such heterogeneity associated with interfacial chemistry can be accounted for when standardization is done within one zone, and herein is a first demonstration for a PAD. In this work, internal standardization and quantitative transduction of nucleic acid hybridization in a multiplexed format was achieved with the gQD/Cy3 and rQD/A647 FRET pairs, respectively. Two-plex hybridization assays were conducted in the absence of mixed films of oligonucleotide probes to achieve a higher assay sensitivity. The sample solutions for the two-plex hybridization assays contained a constant quantity of SMN1 FC TGT at 3.0 pmol, and the amount of uidA FC TGT was varied from 0 pmol to 4.5 pmol. The results for the two-plex hybridization assays are shown in Figure 6. As can be seen from the PL spectra (Figure 6a) and corresponding FRET ratio plot of the two detection channels (Figure 6b), the response from the internal standard channel (gQD/Cy3) remained constant within the precision of the experimental results (relative standard deviation (RSD) < 9%), while the

Table 2. Tuning the Analytical Figures of Merit for the rQD/ A647 FRET Channel by Adjusting the Relative Mole Ratio of the Two Types of QD-Probe Conjugates relative mole ratio of gQD-SMN1 probe to rQD-uidA probe 1:9 1:1 9:1

sensitivity (slope of the response curve)/pmol−1 0.046 (±0.004) 0.078 (±0.014) 0.157 (±0.012)

limit of linearity, (LOL)/pmol (RSD ≤ 16%)

limit of detection (LOD)/ fmol

Figure

3.5

375

5a

2.0

375

5b

0.75

90

5c

a reduction in rQD background was expected to provide improved assay sensitivity and lower LOD. Using this approach, the dynamic range response of the rQD/A647 channel in the two-color assay format was tunable from 90 fmol (30 nM) to 3.5 pmol (1.2 μM). Two-Plex Hybridization Assays in the Two-Color (Multicolor) Format. The use of different FRET pairs for solid-phase QD-FRET multiplexed nucleic acid hybridization assays has been reported in two formats: (1) the use of one FRET pair as an internal standard and the other FRET pair(s) for target sequence determination21 and (2) the use of each of the FRET pairs for the detection of different target sequences.18,20 Internal referencing where the selective chemistries for the internal standard and analyte detection are coimmobilized within one zone is a more reliable approach for

Figure 6. Two-plex hybridization assays with Cy3 labeled SMN1 FC TGT and A647 labeled uidA FC TGT in a multiplexed format, where the gQD/Cy3 FRET pair channel was used as an internal standard channel by keeping the concentration of Cy3 labeled SMN1 FC target constant at 3.0 pmol for all the hybridization assays. The rQD/A647 FRET channel was used as the detection channel. (a) PL spectra for the hybridization of (i) 3.0 pmol/0.0 pmol, (ii) 3.0 pmol/0.75 pmol, (iii) 3.0 pmol/1.5 pmol, (iv) 3.0 pmol/2.2 pmol, (v) 3.0 pmol/3.0 pmol, and (vi) 3.0 pmol/4.5 pmol of SMN1/uidA FC targets. (b) FRET ratio plots for the gQD/Cy3 (green squares) and rQD/uidA (red circles) FRET pairs corresponding to the PL spectra shown in a. (c) Normalized FRET ratio plot corresponding to the FRET ratio plots shown in b. (d) SNP discrimination in the multiplexed format. H

dx.doi.org/10.1021/ac401471n | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

tides with LOD in fmol range and dynamic range spanning 1 order of magnitude. SNP discrimination with a contrast ratio of 70:1 was also possible by using a stringency control that was provided by a combination of ionic strength and formamide. In the two-color format, the control of the conjugation of one type of probe to the corresponding color of QD was shown to provide higher analytical sensitivity (1.5 fold) and dynamic range (1.3 fold) than when probes were indiscriminately conjugated with a mixture of the two colors of QDs. Two-plex hybridization assays were demonstrated by using the gQD/Cy3 FRET pair as an internal standard channel and the rQD/A647 FRET pair as the detection channel. This allowed for an internal standardization to be achieved within each hybridization zone and provided precision for SNP discrimination in the two-color format. The dynamic range response and LOD of the FRET channels were tunable by consideration of the relative concentrations of the two types of QD-probe conjugates. The multiplexed QD-FRET nucleic acid hybridization assays presented in this work can be extended to high density arrays on a paper-based platform for higher-throughput and parallel detection of multiple nucleic acid hybridization reactions, and this work is extendable to the detection of other biomolecules in addition to nucleic acids. Additionally, the hybridization assays presented in this work should also be applicable to the detection of real samples by incorporation of adequate sample processing and nucleic acid amplification steps such as a polymerase chain reaction. Moreover, the methods presented herein are amenable to integration with a fluorescence plate reader to facilitate higher-throughput analysis.15

response from the detection channel (rQD/A647) showed a linear increase for increasing amounts of uidA FC TGT ranging from 0 pmol to 2.2 pmol. In terms of comparing the analytical figures of merit of uidA FC TGT (rQD/A647 FRET channel) in the absence (Figure 4b) and presence of SMN1 FC TGT (Figure 6b), no change in the sensitivities of the response curves was observed. The assay sensitivities in the absence and presence of SMN1 FC TGT were 0.132 (±0.004) pmol−1 and 0.137 (±0.005) pmol−1. These results suggest that the efficiency of hybridization of uidA FC TGT was not compromised in the presence of SMN1 FC TGT in the twoplex assay format. In terms of comparing the LOL of uidA FC TGT (rQD/A647 FRET channel) in the absence (Figure 4b) and presence of SMN1 FC TGT (Figure 6b), the LOL in the absence of SMN1 FC TGT was found to be higher than in the presence of SMN1 FC TGT: 3.0 pmol and ca. 2.2 pmol in the absence and presence of SMN1 FC TGT, respectively. However, the LOL was found to be 3.0 pmol when the FRET ratio response of the rQD/A647 detection channel was normalized to the FRET ratio response of the gQD/Cy3 internal standard channel (normalized FRET ratio, see eq S6 in the Supporting Information) as shown in Figure 6c. This value is consistent with the LOL value of uidA FC TGT in Figure 4b. Hence, the normalized FRET ratio accounts for variations associated with different zones of interfacial chemistry and provided analytical information that was consistent with the earlier results. The LOD of uidA FC TGT in the two-plex assay format was ca. 350 fmol, and this LOD is tunable by controlling the relative concentration of the two types of QD-probe conjugates as discussed earlier. The selectivity of the two-plex hybridization assays was evaluated with NC and FC TGTs (SMN1 and uidA). The presence of SMN1 FC and uidA FC TGTs in a mixture provided a FRET ratio response that was 2 orders of magnitude higher than that for the NC TGTs for the both FRET channels (see Figure S10 in the Supporting Information for results and discussion). SNP detection was also evaluated in the two-plex assay format with sample solutions containing uidA FC TGT in addition to either SMN1 FC TGT or SMN1 1 BPM TGT. The results are shown in Figure 6d. For the corresponding PL images and spectra, see Figure S11 in the Supporting Information. Hybridization assays conducted in BBS buffer yielded a negligible contrast ratio of 1.44:1 for SMN1 FC and SMN1 1 BPM TGTs, respectively. Incubation of the paper in BB buffer containing 10% (v/v) F provided a contrast ratio of ca. 50:1 for SMN1 FC and SMN1 1 BPM TGTs, respectively. These results are consistent with the SNP discrimination results in the single-color assay format. In the presence of BB buffer containing 10% (v/v) F, the FRET ratio responses associated with the SMN1 FC and uidA FC TGTs were retained at 50% and 60%, respectively, of the responses associated with the hybridization conducted in BBS buffer.

■

ASSOCIATED CONTENT

S Supporting Information *

Detailed experimental procedures, description of instrumentation used, characterization of the FRET pairs and equations used in the data analysis, solution-phase hybridization experiments, and additional results and discussion. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors gratefully acknowledge the Natural Sciences and Engineering Research Council of Canada (NSERC) for financial support of this research. M.O.N is also grateful to the Ontario Ministry of Training, Colleges and Universities (MTCU) for provision of an Ontario Graduate Scholarship (OGS).

■

■

CONCLUSIONS The integration of a multiplexed solid-phase QD-FRET nucleic acid hybridization assay on a paper-based platform was investigated. The surface of paper was functionalized with imidazole groups for the immobilization of gQD-SMN1 probe and rQD-uidA probe conjugates that were preassembled in solution. Hybridization assays were demonstrated in the singlecolor (one FRET pair) and two-color (both FRET pairs concurrently implemented) formats. Single-color assays showed excellent resistance to nonspecific adsorption of oligonucleo-

REFERENCES

(1) Algar, W. R.; Susumu, K.; Delehanty, J. B.; Medintz, I. L. Anal. Chem. 2011, 83, 8826−8837. (2) Algar, W. R.; Krull, U. J. Anal. Bioanal. Chem. 2008, 391, 1609− 1618. (3) Algar, W. R.; Tavares, A. J.; Krull, U. J. Anal. Chim. Acta 2010, 673, 1−25. (4) Medintz, I. L.; Mattoussi, H. Phys. Chem. Chem. Phys. 2009, 11, 17−45.

I

dx.doi.org/10.1021/ac401471n | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

(5) Dennis, A. M.; Rhee, W. J.; Sotto, D.; Dublin, S. N.; Bao, G. ACS Nano 2012, 6, 2917−2924. (6) Sapsford, K. E.; Pons, T.; Medintz, I. L.; Mattoussi, H. Sensors 2006, 6, 925−953. (7) Tavares, A. J.; Noor, M. O.; Vannoy, C. H.; Algar, W. R.; Krull, U. J. Anal. Chem. 2012, 84, 312−319. (8) Noor, M. O.; Shahmuradyan, A.; Krull, U. J. Anal. Chem. 2013, 85, 1860−1867. (9) Algar, W. R.; Krull, U. J. Sensors 2011, 11, 6214−6236. (10) Pelton, R. TrAC, Trends Anal. Chem. 2009, 28, 925−942. (11) Li, X.; Ballerini, D. R.; Shen, W. Biomicrofluidics 2012, 6, 11301−1130113. (12) Martinez, A. W.; Phillips, S. T.; Whitesides, G. M.; Carrilho, E. Anal. Chem. 2010, 82, 3−10. (13) Rao, A. N.; Rodesch, C. K.; Grainger, D. W. Anal. Chem. 2012, 84, 9379−9387. (14) Rao, A. N.; Vandencasteele, N.; Gamble, L. J.; Grainger, D. W. Anal. Chem. 2012, 84, 10628−10636. (15) Carrilho, E.; Phillips, S. T.; Vella, S. J.; Martinez, A. W.; Whitesides, G. M. Anal. Chem. 2009, 81, 5990−5998. (16) Cheng, C. M.; Martinez, A. W.; Gong, J. L.; Mace, C. R.; Phillips, S. T.; Carrilho, E.; Mirica, K. A.; Whitesides, G. M. Angew. Chem., Int. Ed. 2010, 49, 4771−4774. (17) Sirvio, J.; Hyvakko, U.; Liimatainen, H.; Niinimaki, J.; Hormi, O. Carbohydr. Polym. 2011, 83, 1293−1297. (18) Algar, W. R.; Krull, U. J. Anal. Chem. 2009, 81, 4113−4120. (19) Algar, W. R.; Krull, U. J. Langmuir 2009, 25, 633−638. (20) Algar, W. R.; Krull, U. J. Langmuir 2010, 26, 6041−6047. (21) Petryayeva, E.; Algar, W. R.; Krull, U. J. Langmuir 2013, 29, 977−987. (22) Noor, M. O.; Tavares, A. J.; Krull, U. J. Anal. Chim. Acta 2013, 788, 148−157. (23) Petryayeva, E.; Krull, U. J. Langmuir 2012, 28, 13943−13951. (24) Araujo, A. C.; Song, Y. J.; Lundeberg, J.; Stahl, P. L.; Brumer, H. Anal. Chem. 2012, 84, 3311−3317. (25) Prasuhn, D. E.; Feltz, A.; Blanco-Canosa, J. B.; Susumu, K.; Stewart, M. H.; Mei, B. C.; Yakovlev, A. V.; Loukov, C.; Mallet, J. M.; Oheim, M.; Dawson, P. E.; Medintz, I. L. ACS Nano 2010, 4, 5487− 5497. (26) Medintz, I. L.; Clapp, A. R.; Brunel, F. M.; Tiefenbrunn, T.; Uyeda, H. T.; Chang, E. L.; Deschamps, J. R.; Dawson, P. E.; Mattoussi, H. Nat. Mater. 2006, 5, 581−589. (27) Algar, W. R.; Krull, U. J. Anal. Chem. 2010, 82, 400−405. (28) Clapp, A. R.; Medintz, I. L.; Mauro, J. M.; Fisher, B. R.; Bawendi, M. G.; Mattoussi, H. J. Am. Chem. Soc. 2004, 126, 301−310. (29) Regulacio, M. D.; Han, M. Y. Acc. Chem. Res. 2010, 43, 621− 630. (30) Algar, W. R.; Krull, U. J. Langmuir 2006, 22, 11346−11352. (31) Algar, W. R.; Krull, U. J. J. Colloid Interface Sci. 2011, 359, 148− 154. (32) Gong, P.; Levicky, R. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 5301−5306. (33) Watterson, J. H.; Piunno, P. A. E.; Wust, C. C.; Krull, U. J. Langmuir 2000, 16, 4984−4992. (34) Sadhu, C.; Dutta, S.; Gopinathan, K. P. J. Biosci. 1984, 6, 817− 821. (35) Liu, T. C.; Huang, Z. L.; Wang, H. Q.; Wang, J. H.; Li, X. Q.; Zhao, Y. D.; Luo, Q. M. Anal. Chim. Acta 2006, 559, 120−123. (36) Clapp, A. R.; Medintz, I. L.; Uyeda, H. T.; Fisher, B. R.; Goldman, E. R.; Bawendi, M. G.; Mattoussi, H. J. Am. Chem. Soc. 2005, 127, 18212−18221.

J

dx.doi.org/10.1021/ac401471n | Anal. Chem. XXXX, XXX, XXX−XXX