Exponential Amplification for Chemiluminescence Resonance Energy

Mar 29, 2011 - The human lung total RNA (1 μg/μL) and human lung tumor total RNA (1 μg/μL) were purchased from Agilent (Santa Clara, CA) and Takar...
0 downloads 15 Views 2MB Size
ARTICLE pubs.acs.org/ac

Exponential Amplification for Chemiluminescence Resonance Energy Transfer Detection of MicroRNA in Real Samples Based on a Cross-Catalyst Strand-Displacement Network Sai Bi, Jilei Zhang, Shuangyuan Hao, Caifeng Ding, and Shusheng Zhang* State Key Laboratory Base of Eco-chemical Engineering, College of Chemistry and Molecular Engineering, Qingdao University of Science and Technology, Qingdao 266042, People's Republic of China

bS Supporting Information ABSTRACT: An exponential amplification strategy for ultrasensitive detection of microRNA (miRNA) in biological extracts is developed based on a cross-catalyst stranddisplacement reaction (CC-SDR). Functionally, the system consists of one upstream circuit and two downstream circuits, each of which comprises a three-stranded substrate complex and a single-stranded fuel. Importantly, the exponential amplification process does not require a polymerase or a nicking endonuclease. The whole network is activated by a miRNA trigger in the upstream circuit, which regenerates the miRNA trigger to catalyze another new upstream circuit and release two specified DNA outputs to further act as catalysts of the two downstream circuits, respectively. During each cross-catalyst network, two “mimic trigger” DNA are generated, which in turn catalyze the upstream system. Finally, the exponentially produced luminol-reduced AuNPs (lumAuNPs) and fluorescein-tagged signals are sensitively read out in the form of luminolH2O2horseradish peroxide (HRP)fluorescein chemiluminescence resonance energy transfer (CRET) triplex probes by employing magnetic nanoparticles to reduce high background, achieving a detection limit of let-7a miRNA as low as 0.68 fM. Moreover, the proposed strategy exhibits an excellent specificity to discriminate one-base differences among the let-7 miRNA family and is successfully applied in real sample assay: let-7a miRNA in total RNA samples extracted from human lung tissue, and let-7b miRNA from human lung cancer cells and cervical adenocarcinoma cells, respectively. To the best of our knowledge, this is the first study to use the chemiluminescence technique for miRNA detection, which can be expected to provide a new and ultrasensitive platform for amplified detection and subsequent analysis of miRNA.

A

class of endogenous, non-protein-coding small RNAs called microRNAs (miRNAs) (approximately 1824 nucleotides in length) has attracted increasing scientific interest due to their potent regulatory role in gene expression.1,2 Recent studies have revealed that specific alteration in miRNA expression profiles are directly associated with various diseases, such as human cancers, cardiovascular diseases, and viral infections.35 Therefore, miRNAs could be considered as a family of clinically important biomarkers for early diagnosis and pathogenesis of diseases, therapeutic intervention, and use in basic biomedical research.6,7 However, the intrinsic properties of miRNAs, such as short size, sequence similarity among family members, low abundance, and susceptibility to degradation, make the accurate detection and quantitation more difficult and challenging. So far, several analytical approaches have been developed for miRNA expression profiling analysis.8 Northern blotting methodology is regarded as a standard method for miRNA detection but is often restricted by problems of comparative ponderosity, demand for a large amount of miRNAs, and semiquantitative analysis.9 Microassay-based techniques offer a simple and highthroughput platform for miRNA expression profiling. Unfortunately, the sensitivity and specificity are still not satisfactory.10 In order to achieve the detection of miRNAs at lower expression r 2011 American Chemical Society

levels, various sensitive strategies have been developed, such as modification of the Invader assay,11 signal-amplifying ribozymes,12 rolling circle amplification,13,14 and nanoparticlebased amplification.15,16 Especially, real-time quantitative PCR (RT-qPCR) has become a standard approach for simultaneous amplification and quantification of miRNA. For example, the stem-loop reverse transcription primer method that was developed by Chen et al. has achieved a highly specific and ultrasensitive PCR-based quantification of miRNAs as low as the single molecule level.17 Lately, an isothermally exponential amplification reaction (EXPAR) has been introduced into ultrasensitive detection of nucleic acids.1820 However, the utilization of enzymes (e.g., polymerase and nicking endonuclease) would certainly restrict the universal application of the method for nucleic acid detection because during the reaction the amount and activity of the enzymes have to be considered, and a specific DNA sequence for the nicking endonuclease recognition site is also required. Thus, the development of an enzyme-free method Received: January 13, 2011 Accepted: March 29, 2011 Published: March 29, 2011 3696

dx.doi.org/10.1021/ac200096b | Anal. Chem. 2011, 83, 3696–3702

Analytical Chemistry

ARTICLE

Scheme 1. Schematic Representation of miRNA Detection through Enzyme-Free Exponential Amplification Strategy Based on the Proposed CC-SDR Process, Which Is Triggered and Catalyzed by the miRNA Trigger (target) of the Upstream Circuit and Autocatalyzed by the DNA Trigger Generated from the Downstream Circuita

a The output signals (S1 and S2) are sensitively read out in the form of the CRET triplex probe by employing MNPs to reduce the high background induced by lumAuNP- and fluorescein-labeled signals sequestered on the duplex complexes (C1 and C2).

for miRNA detection would gain great significance that could provide an applicable approach for clinical assays. Recently, the strand-displacement reaction (SDR) has become increasingly attractive for the development of molecular nanotechnology.21,22 SDR is a process of executing a conditional displacement event through binding a full-length complementary strand to a single-stranded overhang referred to as a “toehold” that is initially sequestered in a stable duplex pattern, which is further utilized to invade and displace a hybridized sequence via a strandexchange mechanism of three-way branch migration.2325 In light of the specific and fast strand recognition, various molecular machines have been developed,2628 and even logical control elements and circuits have been constructed.2931 In comparison with other strand disassembly methods by using temperature,32 enzymes,33 and apatamers,34 one unique advantage of SDR is that the reaction mechanism presented here is based on branch migration and driven forward thermodynamically by the entropy from liberated molecules without enzymatic catalysis.35 Therefore, SDR holds great promise as a competent method for miRNA detection through contriving amplification strategy to improve sensitivity. Herein we introduce a cross-catalyst SDR (CC-SDR) strategy into exponentially amplified detection of a member of the miRNA family, lethal-7 (let-7), one of the well-known biomarkers to lung cancer,36 extracted from human lung tissue and cancer cell lines (carcinoma of lung and cervix). As shown in Scheme 1, the whole network is divided into two layers: one upstream circuit and two downstream circuits. Each of them consists of a three-stranded substrate complex (C) and a fuel (F). In brief, the let-7a miRNA trigger is first employed as both input and catalyst to trigger the upstream circuit, resulting in the release of two specified DNA outputs (O1 and O2) and regeneration of the miRNA trigger to trigger another new upstream circuit. Then the two released outputs (O1 and O2) further act as catalysts for two downstream circuits, respectively, producing free singlestranded DNA triggers in turn to serve as the “mimic target” to catalyze the original upstream system. The reaction is then autocatalytic and cross-catalytic. During this two-layer crosscatalytic feedback circuit, although the concentration of the upstream catalysts (miRNA triggers) is constant, the concentration of released DNA triggers increases exponentially with time,

Scheme 2. Schematic Diagram of the Flow Injection System for miRNA Detection

which facilitates the exponential amplification for miRNA detection. Furthermore, the exponentially produced signals, S1 tagged with luminol-reduced AuNPs (lumAuNPs) and S2 tagged with fluorescein, are hybridized with HRPDNA on magnetic nanoparticles (MNPs) with the assistance of DNA binder BePI into triplex probes to perform the luminolH2O2HRPfluorescein chemiluminescence resonance energy transfer (CRET) system37 using flow-injection detection. The flow injection mode for the miRNA assay is shown in Scheme 2. In addition, MNPs employed here play an important role not only in the simplification of the manipulation but also for the reduction of high CRET background to improve the detection sensitivity (vide infra). To the best of our knowledge, this is the first report using the chemiluminescence technique for miRNA detection, even though methods of surface-enhanced Raman spectroscopy,6 fluorescence,1420 and electrochemistry38,39 have already been studied.

’ EXPERIMENTAL SECTION Chemicals and Apparatus. miRNAs and other oligonucleotides used in this study were synthesized and purified by Takara Biotechnology Co., Ltd. (Dalian, China) and Sangon Biotechnology Co., Ltd. (Shanghai, China), respectively (for sequences of miRNA and DNA, see Table S1 in Supporting Information). The RNA oligonucleotides were purified by HPLC, which were first prepared as 1.0  104 M stock solution, followed by diluting to desirable concentration by a stepwise dilution method to ensure that the miRNAs maintain high quality and to improve 3697

dx.doi.org/10.1021/ac200096b |Anal. Chem. 2011, 83, 3696–3702

Analytical Chemistry the accuracy of sample concentration.14 Benzo[e]pyridoindole (1,2-benzpyrene or 4,5-benzopyrene) (BePI) was ordered from Acros Organics (Geel, Belgium). Carboxyl group-modified magnetic nanoparticles (MNPs) (0.51.0 μm, 10 mg/mL) were commercially available from BaseLine ChromTech Research Centre (Tianjin, China). StreptavidinHRP was purchased from Boster Bio-Engineering Co., Ltd. (Wuhan, China). Tris(2-carboxyethyl)phosphine hydrochloride (TCEP, 98%) was obtained from Alfa Aesar (Ward Hill, MA). Luminol standard powder, 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydro-chloride (EDC), and N-hydroxysuccinimide (NHS) were ordered from Sigma-Aldrich. HAuCl4 3 4H2O was ordered from Shanghai Reagent Co., Ltd. (Shanghai, China). A luminol stock solution (1.0  102 M) was prepared by dissolving in 0.1 M NaOH solution and further stored in the dark. A HAuCl4 stock solution (1%, w/w) was prepared by dissolving 1 g of HAuCl4 in 100 mL of doubly distilled deionized water and stored at 4 °C for further use. The human lung total RNA (1 μg/μL) and human lung tumor total RNA (1 μg/μL) were purchased from Agilent (Santa Clara, CA) and Takara (Clontech, Mountain View, CA), respectively. Doubly distilled deionized water was used throughout the experiments. All the reagents were of analytical grade and used without further purification. The CRET measurements were performed with a FI-CL instrument (MPI-F, Remex Analytical Instrument Co. Ltd., Xi’an, China). Transmission electron microscopy (TEM) images were taken on a JEOL JEM-2000EX. UVvisible spectra were acquired with a Cary 50 UVvisNIR spectrophotometer (Varian), and fluorescence spectra were measured on a F-4500 fluorescence spectrophotometer (Hitachi, Japan). The RT-qPCR was performed on an MX3000p (Stratagene). Synthesis of LumAuNPs. LumAuNPs were synthesized according to the method by reducing tetrachloroauric acid with luminol.30,40 Briefly, 100 mL of 0.02% (w/w) HAuCl4 solution was heated to boiling with vigorous stirring, followed by adding 1.8 mL of 1.0  102 M luminol stock solution rapidly. After continued boiling for 40 min, the color of the solution turned from yellow to black to purple and finally reached a wine-red color, indicating the formation of lumAuNPs. After the resulting colloidal suspension was cooled to room temperature with stirring, dialysis was carried out to remove excess luminol from the AuNPs. The synthesized lumAuNPs were characterized by UVvisible spectroscopy, fluorescence spectroscopy, and TEM, respectively (see Supporting Information). Preparation of LumAuNP-Labeled Oligonucleotides. The lumAuNP-labeled DNA (lumAuNP-DNA, S1) was fabricated by adding ∼500 μL of the as-prepared lumAuNPs to 100 μL of 1.0  107 M thiol-modified DNA which was activated with TCEP (10 mM) for 1 h prior to attachment to lumAuNPs. The synthesized lumAuNPDNA conjugates were shaken gently for 16 h at room temperature and “aged” in the solution (0.3 M NaCl, 10 mM Tris-acetate, pH 8.2) for another 48 h. Following removal of excess reagents by centrifugation at 10 000 rpm for 10 min, the resulting red oily precipitate was washed with 500 μL of 0.01 M PBS buffer containing 0.01 M NaCl (pH 7.4), recentrifuged, redispersed in 500 μL of the same buffer, and stored at 4 °C for further use. Preparation of HRPDNA-Labeled MNPs (H-DNA/MNPs). First, a 30 μL suspension of carboxylated MNPs was washed three times with 200 μL of 0.01 M PBS buffer (pH 7.4), followed by adding 120 μL of 0.1 M imidazoleHCl buffer containing 0.2 M EDC and incubating at 37 °C for 30 min to activate the

ARTICLE

carboxylate groups on the MNPs. After washing three times with 200 μL of PBS buffer, 100 μL of 1.0  107 M 50 -biotin-modified and 30 -amino-modifed DNA was added and reacted with the activated carboxyl groups of MNPs at 37 °C overnight. Then 100 μL of 1.0  107 M streptavidinHRP was added and incubated for 40 min at 37 °C to obtain HRPDNA-labeled MNPs (H-DNA/MNPs) via biotinstreptavidin specific recognition, followed by washing three times with 200 μL of 0.01 M PBS at room temperature, resuspending in 2 mL of PBS buffer, and storing at 4 °C for subsequent use. The TEM images of the carboxylated MNPs before and after conjugation with HRPDNA were recorded as shown in Figure S3 in Supporting Information. Nondenaturing Polyacrylamide Gel Electrophoresis. The samples of both the upstream circuit and the whole two-layer network after triggering by target let-7a miRNA at 0 and 2 h were characterized by 30% native polyacrylamide gel electrophoresis (Acr = acrylamide, Bis = N,N0 -methylenebisacrylamide; Acr/ Bis = 29/1), respectively. Tris-acetateEDTA (TAE) (pH = 8.5) was used as the separation buffer. Electrophoresis was carried out at 120 V for 1 h at 25 °C. miRNA Detection. All three-stranded substrate complexes (C0, C1, and C2) were previously hybridized by mixing each corresponding oligonucleotide at the same concentration of 1.0  108 M at 25 °C for 30 min. Subsequently, 200 μL of miRNA target at different concentrations was added to the substrate solution containing the fabricated C0, C1, C2 complexes, and F0, F1, F2 sequences (all at a concentration of 1.0  108 M), followed by reacting at 25 °C in Tris-acetate buffer containing 12.5 mM MgCl2 for 2 h.27 Then 300 μL of HRP DNA/MNPs and 10 μL of DNA triplex molecular binder BePI (0.1 mM) were added and incubated at 25 °C for 1 h to form the triplexMNP conjugates. After magnetic separation, the resulting triplexMNP products were washed twice with 0.01 M PBS buffer containing 0.01 M NaCl (pH 7.4) and redispersed in 2 mL of the same PBS buffer, followed by FI-CL detection. FI-CL Detection. The CRET measurements were performed with a FI-CL instrument (MPI-F, Remex Analytical Instrument Co. Ltd., Xi’an, China) that consisted of a flow injection system (IFIS-D), a chemiluminescence detector (RFL-1), and a computer. The IFIS-D injection system contained two peristaltic pumps which were used to deliver the sample stream (assistant pump) and H2O2 stream (main pump), respectively. Polytetrafluoroethylene tube was used to connect all the components of the flow system. First, the sample containing the resulting product of the triplexMNPs was pumped. Then the valve turned to water injection, and the triplexMNPs were moved forward by water, mixed with H2O2, and then reached the twisty flow cell. The total emitted chemiluminescence was collected with a photomultiplier tube at a voltage of 500 V, and the signals were recorded using a computer. The relative peak height of total FI-CL intensity (deduction of blank) was measured to construct the calibration curve versus the concentration of target miRNA. Cell Lysis and RNA Preparation. HeLa cells were cultured according to the instructions of American Type Culture Collection. The cell line was grown in RPMI 1640 medium supplemented with 10% fetal bovine serum and 100 IU/mL penicillinstreptomycin at 37 °C, followed by culturing in a humidified atmosphere of 95% air and 5% CO2. The density of the cells was counted with a hemocytometer. A 1.0 mL amount of ∼6  106 suspended cells was centrifuged at 3500 rpm for 5 min 3698

dx.doi.org/10.1021/ac200096b |Anal. Chem. 2011, 83, 3696–3702

Analytical Chemistry in culture medium, washed once with PBS buffer, and spun down at 3500 rpm for 5 min. Then the cell pellets were suspended in 600 μL of lysis solution. The total RNA extracted from HeLa cells was prepared by using RNA Prep Pure Cell Kit (Tiangen Biotechnology Co., Ltd., Beijing, China) according to the manufacturer’s procedures. The RNA concentration was determined to be 2.14 pg/cell from the UVvis absorption at 260 nm.

’ RESULTS AND DISCUSSION Principle of the Proposed Exponential Amplification for miRNA Detection. First, note the upstream circuit, for example,

to illustrate the performance of the proposed two-layer crosscatalytic feedback circuit for exponentially amplified detection of let-7a miRNA (Scheme 1). Preparatively, output 1 (O1), output 2 (O2), and linker 0 (L0) are fabricated from a three-stranded substrate complex (C0) with a toehold domain on L0 and inert domains on O1 and O2, respectively. Upon the introduction of the miRNA trigger, the miRNA at its 30 terminus hybridizes with the single-stranded toehold domain of L0 at its 50 terminus to form a four-stranded intermediate 1 (I1) and then quickly rearranges into intermediate 2 (I2) by branch migration, enabling O2 to possess a unique toehold. In this I2 state, the binding between O2 and L0 is too weak to maintain the O2 attachment to L0, so O2 spontaneously dissociates from I2 which further yields a free strand O2 and transfers to intermediate 3 (I3) with a newly exposed single-stranded domain facilitating the hybridization of single-stranded fuel 0 (F0) to form intermediate 4 (I4). The state of I4 is unstable, which quickly transfers to release O1 and intermediate 5 (I5). On the basis of branch migration and driven by entropy, F0 finally displaces the miRNA trigger so that I5 rearranges to spontaneously dissociate as double-stranded waste (W1) and rerelease the miRNA trigger as catalyst to implement another circulation of the upstream circuit. In summary of the whole course of the reaction, the net reaction of one cycle of the upstream circuit could be clarified as a single-stranded fuel that reacts with a three-stranded substrate complex, displacing two kinds of output from the linker strand of the substrate to form a double-stranded waste, while the miRNA trigger is rereleased to catalyze another cycle of the upstream circuit. The reaction can be repeated continuously, resulting in the linear amplification of the target miRNA. Subsequently, the released outputs O1 and O2 of the upstream circuit alternatively serve as the catalysts to trigger two downstream circuits, respectively, releasing S1 and S2 signals. Moreover, the released DNA triggers in the downstream circuits serve as targets to in turn autocatalyze the upstream circuit and repeat the above cyclic process to implement the CCSDR operation. The sequence of the released DNA trigger is the same as that of the miRNA target, except that the ribonucleotides and uridine in the miRNA are respectively substituted by deoxyribo-nucleotides and thymine in the DNA strand. From the previous study, the melting temperatures (Tm) of let-7a miRNA and its same-sequence DNA were determined to be 64.7 and 67.2 °C, respectively.19 Because the CC-SDR is performed at 25 °C, the reaction could be performed as designed. Hybridization of these released DNA triggers with C0 of the upstream circuit leads to automatic operation of the CC-SDR network, leading to an exponential amplification of the target miRNA. After the accomplishment of the two-layer network, HRPDNA-labeled MNPs and DNA binding molecules BePI are added and bound with the released S1 and S2 into triplex probes to generate a CRET signal. The mechanism of the luminolH2O2HRPfluorescein

ARTICLE

CRET system was extensively investigated in this study, and the chemiluminescence spectra of the system both in free state and binding state were also recorded (see Supporting Information for details). The triplex binder BePI used here has another important function in terminating the CC-SDR process by stabilizing the unreacted duplex substrate complex in solution. Note that every component of the whole network has been fully utilized without any “idle” single strand. For example, some of the outputs of the two downstream circuits (DNA trigger) in turn act as the new catalysts of the upstream, while the others ingeniously serve as the signals (S1 and S2) to bind with HRPDNA on MNPs into triplex probes to facilitate CRET detection. In addition, the design of the strategy is straightforward and flexible: miRNA trigger, DNA trigger, and O1 are entirely independent in sequence, while the sequences of other strands could be changed accordingly with the new analyte. Nondenaturing Polyacrylamide Gel Electrophoresis. In this CC-SDR system, the triggers, fuels, outputs, and signals are all single-stranded with similar lengths; each molecule can play multiple roles within the network. For example, O1 and O2 simultaneously act as outputs of the upstream circuit and catalysts of the downstream circuit. Here, nondenaturing polyacrylamide gel electrophoresis (PAGE) was used to verify the catalytic pathway of only the upstream circuit (Figure 1A) and autocatalytic pathway of the whole two-layer network (Figure 1B). Both processes after triggering by the let-7a miRNA target were conducted and analyzed on PAGE at 0 h (lane 1 for upstream circuit and lane 3 for the whole two-layer network) and 2 h (lane 2 for upstream circuit and lane 4 for the whole two-layer network), respectively. The wide bands of lanes 3 and 4 could be attributed to the multiple components of the whole network (e.g., corresponding C, W, and intermediate states of one upstream and two downstreams). From the results, the total number of base pairs in the reactants and products of not only the upstream circuit but also the whole two-layer network could be considered as almost unchanged, which further confirmed the feasibility of the proposed assay. Sensitivity of the Assay. The let-7a miRNA are subsequently detected by the proposed strategy to evaluate the sensitivity and range over which the assay is linear. After optimization of the concentrations of H2O2 (0.75 mM) and initial reactants (1.0  108 M), and the amount of HRPDNA-labeled MNPs (300 μL) used in this study (see Supporting Information), we adopted the concentration of target corresponding to the peak height of the flow-injection chemiluminescence (FI-CL) intensity curve as calibration standard for quantitative detection of let-7a miRNA. From Figure 2, as the target concentration increases from 1.0 to 100.0 fM, the FI-CL signal from the fabricated two-layer cross-catalytic feedback network increases significantly. There is a fairly good linear relationship between the relative FI-CL intensity and the concentration of let-7a miRNA in the range of 1.010.0 fM. The regression equation is expressed as I = 29.6800C þ 4.8456 (I represents the peak height of FI-CL; C represents the concentration of let-7a miRNA, fM; n = 7, correlation coefficient R = 0.9982) with a limit of detection (LOD) of 0.68 fM let-7a miRNA (3σ), which is approximately 3 orders of magnitude lower than that obtained by only executing the upstream circuit for a linear amplification (see Supporting Information). In comparison with other methods for miRNA assay, such as RT-qPCR or microarrays, the detection limit of the proposed strategy is relatively insensitive. However, as a PCR-free technology, the proposed assay has achieved one of the most sensitive approaches for miRNA detection compared to previous PCR-free-based methods (Table S1 in Supporting 3699

dx.doi.org/10.1021/ac200096b |Anal. Chem. 2011, 83, 3696–3702

Analytical Chemistry

ARTICLE

Figure 1. Nondenaturing PAGE of the upstream circuit (A) and the whole two-layer network (B). M: marker; lanes 1 and 2: the upstream circuit after the introduction of let-7a miRNA target at 0 and 2 h, respectively; lanes 3 and 4: the whole two-layer network after the introduction of let-7a miRNA target at 0 and 2 h, respectively.

Figure 3. The comparison of FI-CL signals for the blank samples (in the absence of target let-7a miRNA) and 10.0 fM let-7a miRNA samples by employing MNPs (curve a: blank; curve c: 10.0 fM let-7a miRNA) and without employing MNPs (curve b: blank; curve d: 10.0 fM let-7a miRNA).

Figure 2. (A) FI-CL kinetic curve for CRET intensity generated by Scheme 1 corresponding to different concentrations of let-7a miRNA: (a) 0; (b) 1.0; (c) 2.0; (d) 4.0; (e) 6.0; (f) 8.0; (g) 10.0; (h) 20.0; (i) 40.0; (j) 60.0; (k) 80.0; (l) 100.0 fM. (B) The corresponding calibration curve of peak height versus the concentration of let-7a miRNA. Inset: magnification of the plot in the range of 010.0 fM let-7a miRNA.

Information). A relative standard deviation (RSD) of 7.2% for 11 repetitive measurements of 40.0 fM let-7a is obtained, providing a

good reproducibility of this miRNA assay. As a tumor suppressor, it has been found that let-7 miRNA is poorly expressed in carcinoma of the lung.36 Thus, the proposed ultrasensitive method for let-7a miRNA detection satisfies the accurate quantitation of miRNA even at low concentration, which is of great significance in the early diagnosis and effective therapy of cancers. In addition, a control experiment was carried out without employing MNPs to reduce the CRET background. From Figure 3, in this case for the detection of blank sample (in the absence of target), a considerably high FI-CL signal was detected (∼262, curve b), which was much higher than that employing MNPs to immobilize the triplex probes (∼31, curve a). For the detection of 10.0 fM let-7a miRNA, an S/B (the signal ratio of sample to blank) of 10.7 and 2.0 was obtained by employing MNPs and without employing MNPs, respectively. The results 3700

dx.doi.org/10.1021/ac200096b |Anal. Chem. 2011, 83, 3696–3702

Analytical Chemistry

Figure 4. FI-CL signals for CRET intensity as shown in Scheme 1 triggered by let-7ag and let-7i miRNA (all at a concentration of 40.0 fM). Each value is the average of three measurements.

indicated that serious background interference occurred during the reaction, which would definitely affect the detection sensitivity. The high background induced in this system can be ascribed that a random FI-CL signal would occur in this luminol H2O2HRPfluorescein CRET system resulting from lumAuNP DNA and fluoresceinDNA signals, respectively, sequestered on their substrate complexes (C1 and C2) but not the triplex probes, even though the distances between the two signals did not facilitate the CRET reaction in homogeneous solution. From the results, the LOD of let-7a miRNA under homogeneous reaction is approximately 1 order of magnitude higher than that employing MNPs to reduce background. In order to realize the highest sensitivity of the exponential amplification strategy, it is quite necessary to employ MNPs to overcome the high CRET background in this study (see Supporting Information). Therefore, the ultrahigh sensitivity achieved in the proposed method is attributed not only to the CC-SDR but also to the successful conquest of the high background inherent in the CRET system. Moreover, in comparison with other detection tools, especially fluorescence detection, the FI-CL technique exhibits higher sensitivity and better signal resolution, which has certainly made a indubitable contribution to the ultralow LOD. Specificity Study. A significant challenge for miRNA analysis is the ability to distinguish many highly similar sequences that differ by only a few nucleotides. Although most human miRNAs differ in four or more nucleotide bases, members of the let-7 miRNA family with high sequence homology differ by only one or two bases in sequences with the same length.36 Thus, the let-7 miRNA family (let-7ag,i) are used as a model system to evaluate the specificity of the proposed miRNA assay. As shown in Figure 4, the FI-CL signal generated by let-7a miRNA could obviously be distinguished from those produced by other let-7 miRNAs. Thus, the proposed miRNA assay with exponential amplification exhibits an excellent discrimination of all let-7a miRNA family members, even though a single base is mismatched. The high specificity of this protocol could be attributed to the miRNA at the 30 terminus to initiate the toehold exchange reaction. Considering that the mismatched bases in let-7bd,g,i are located near their 30 terminus relative to let-7a; it is thus credible that these mismatched let-7 miRNAs could efficiently be distinguished from let-7a miRNAs. For let-e,f, although their

ARTICLE

mismatched bases are relatively distant from their 30 terminus, they could still remarkably discriminate from let-7a owing to the proposed assay based on the principle of strand recognition with high specificity. Real Sample Assay. A key advantage of the enzyme-free exponential amplification strategy for miRNA detection is its potential for real sample analysis. Given that the expression levels of the let-7 miRNA family are closely associated with cell development and human expression profiles,36 first we apply the proposed assay for let-7a miRNA analysis of total RNA extracted from normal human lung tissue. A series of synthetic let-7a miRNAs at concentrations of 0, 1.0, 2.0, 4.0, 6.0, 8.0, and 10.0 fM were spiked into 1.0 pg/μL diluted human lung total RNA samples, respectively, with equal volume to establish a calibration curve, and the content of let-7a in 1.0 μg/μL human lung total RNA sample was estimated to be 9.76 nM (see Supporting Information). In addition, the reliability of the method was confirmed by adding 10.0 nM synthetic let-7a miRNA to 1.0 μg/μL total RNA sample followed by five replicate measurements. The average content of let-7a miRNA was determined to be 19.81 nM. Using the detection limit of 0.68 fM let-7a miRNA demonstrated above, the proposed method holds great potential for the quantitation of miRNA in real total RNA sample analysis. Moreover, we used the method for let-7b miRNA analysis in the total RNA extracted from cancer cells, demonstrating not only the applicability of the strategy for miRNA detection in real world samples but also the feasibility of the method just by changing the sequences of L0, L1, L2, F1, F2, and the DNA trigger according to the target (for these sequences, see Table S1). First, we analyzed the content of let-7b miRNA in commercial human lung tumor total RNA samples. The diluted total RNA was directly applied to the proposed assay. From the results, the concentration of let-7b miRNA in the total RNA extracted from lung cancer cells is (1.92 ( 0.28)  107 copies/μg RNA, which is consistent with the reported results of miRNA expression profiling.41,42 An RSD < 10% for 11 measurements of total RNA extracted from lung cancer cells offers satisfactory accuracy for the assay in the identification of miRNA expression with slight differences and an excellent specificity, excluding the interference of other species. Subsequently, we analyzed the content of let-7b in cervical adenocarcinoma (HeLa cells) total RNA. The total RNA sample was prepared according to a reported procedure.12 The concentration of let-7b in total RNA extracted from HeLa cells is calculated to be (2.08 ( 0.25)  107 copies/μg RNA.41,42 Considering the detection limit of the assay, it is promising that a dozen cells were able to provide an adequate amount of total RNA for miRNA detection. For comparison, RT-qPCR was conducted for the analysis of the same sample (the details for RT-qPCR detection are described in Supporting Information). The comparative results from these two methods showed relatively good correlations with a correlation coefficient of 0.9997.

’ CONCLUSIONS In summary, an exponentially amplified and highly selective detection of label-free microRNAs is accomplished based on a two-layer CC-SDR network, which is initiated and catalyzed by an miRNA target of the upstream circuit and further autocatalyzed by a DNA trigger of the downstream circuit. In comparison with previously reported methods, the whole exponential amplification 3701

dx.doi.org/10.1021/ac200096b |Anal. Chem. 2011, 83, 3696–3702

Analytical Chemistry process does not require any enzyme, such as a polymerase or a nicking endonuclease. The generated signals can be sensitively read out in the form of CRET triplex probes through flow injection measurement. As far as we know, this is the first report of a chemiluminescence technique for miRNA detection. Moreover, MNPs are employed here as carriers of CRET triplex probes to reduce the high CRET background and simplify the manipulation, achieving a detection limit of let-7a miRNA as low as 0.68 fM. In addition, the proposed strategy has successfully achieved the detection of human miRNAs from tissue- and cancer cell-specific total RNA extracts. As a simple, low-cost, and ultrasensitive method for the detection of miRNA expression levels, the proposed assay holds great promise for considerably advancing the field of routine gene expression profiling and clinical diagnostics, particularly for early cancer diagnosis and designed drug therapy.

’ ASSOCIATED CONTENT

bS

Supporting Information. Additional information as noted in the text. The material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*Phone: þ86-532-84022750. Fax: þ86-532-84022750. E-mail: [email protected].

’ ACKNOWLEDGMENT This work was supported by National Natural Science Foundation of China (21025523), the Basic Research Project of Qingdao (10-3-4-4-5-jch), and the National Basic Research Program of China (2010CB732404). ’ REFERENCES (1) Lee, R. C.; Feinbaum, R. L.; Ambos, V. Cell 1993, 75, 843–854. (2) Lu, J.; Getz, G.; Miska, E. A.; Alvarez-Saavedra, E.; Lamb, J.; Peck, D.; Sweet-Cordero, A.; Ebert, B. L.; Mak, R. H.; Ferrando, A. A.; Downing, J. R.; Jacks, T.; Horvitz, H. R.; Golub, T. R. Nature 2005, 435, 834–838. (3) Calin, G. A.; Croce, C. M. Nat. Rev. Cancer 2006, 6, 857–866. (4) Zhang, C. Clin. Sci. 2008, 114, 699–706. (5) Sullivan, C. S.; Ganem, D. Mol. Cell 2005, 20, 3–7. (6) Arenz, C. Angew. Chem., Int. Ed. 2006, 45, 5048–5050. (7) Cissell, K. A.; Rahimi, Y.; Shrestha, S.; Hunt, E. A.; Deo, S. K. Anal. Chem. 2008, 80, 2319–2325. (8) Cissell, K. A.; Deo, S. K. Anal. Bioanal. Chem. 2009, 394, 1109–1116. (9) Varallyay, E.; Burgyan, J.; Havelda, Z. Nat. Protocols 2008, 3, 190–196. (10) Lee, J. M.; Cho, H.; Jung, Y. Angew. Chem., Int. Ed. 2010, 49, 8662–8665. (11) Allawi, H. T.; Dahlberg, J. E.; Olson, S.; Lund, E.; Olson, M.; Ma, W.-P.; Takova, T.; Neri, B. P.; Lyamichev, V. I. RNA 2004, 10, 1153–1161. (12) Hartig, J. S.; Gr€une, I.; Hajafi-Shoushtari, S. H.; Famulok, M. J. Am. Chem. Soc. 2004, 126, 722–723. (13) Jonstrup, S. P.; Koch, J.; Kjems, J. RNA 2006, 12, 1747–1752. (14) Cheng, Y.; Zhang, X.; Li, Z.; Jiao, X.; Wang, Y.; Zhang, Y. Angew. Chem., Int. Ed. 2009, 48, 3268–3272. (15) Li, J.; Schachermeyer, S.; Wang, Y.; Yin, Y.; Zhong, W. Anal. Chem. 2009, 81, 9723–9729. (16) Zhang, J.; Fu, Y.; Mei, Y.; Jiang, F.; Lakowicz, J. R. Anal. Chem. 2010, 82, 4464–4471.

ARTICLE

(17) Chen, C.; Ridzon, D. A.; Broomer, A. J.; Zhou, Z.; Lee, D. H.; Nguyen, J. T.; Barbisin, M.; Xu, N. L.; Mahuvakar, V. R.; Andersen, M. R.; Lao, K. Q.; Livak, K. J.; Guegler, K. J. Nucleic Acids Res. 2005, 33, e179. (18) Connolly, A. R.; Trau, M. Angew. Chem., Int. Ed. 2010, 49, 2720–2723. (19) Jia, H.; Li, Z.; Liu, C.; Cheng, Y. Angew. Chem., Int. Ed. 2010, 49, 5498–5501. (20) Tan, E.; Wong, J.; Nguyen, D.; Xhang, Y.; Erwin, B.; Van Ness, L. K.; Baker, S. M.; Galas, D. J.; Niemz, A. Anal. Chem. 2005, 77, 7984–7992. (21) Bath, J.; Turberfield, A. J. Nat. Nanotechnol. 2007, 2, 275–284. (22) Zhang, D. Y. J. Am. Chem. Soc. 2011, 133 (4), 1077–1086 . (23) Yurke, B.; Turberfield, A. J.; Mills, A. P., Jr.; Simmel, F. C.; Neumann, J. L. Nature 2000, 406, 605–608. (24) Zhang, Z.; Cheng, Q.; Feng, P. Angew. Chem., Int. Ed. 2009, 48, 118–122. (25) Zhang, D. Y.; Winfree, E. J. Am. Chem. Soc. 2009, 131, 17303–17314. (26) Lubrich, D.; Lin, J.; Yan, J. Angew. Chem., Int. Ed. 2008, 47, 7026–7028. (27) Omabegho, T.; Sha, R.; Seeman, N. C. Science 2009, 324, 67–71. (28) Lubrich, D.; Green, S. J.; Turberfield, A. J. J. Am. Chem. Soc. 2009, 131, 2422–2423. (29) Frezza, B. M.; Cockroft, S. L.; Ghadiri, M. R. J. Am. Chem. Soc. 2007, 129, 14875–14879. (30) Zhang, D. Y.; Turberfield, A. J.; Yurke, B.; Winfree, E. Science 2007, 318, 1121–1125. (31) Soloveichik, D.; Seelig, G.; Winfree, E. Proc. Natl. Acad. Sci. U.S. A. 2010, 107, 5393–5398. (32) Huang, Y.; Zhang, Y.-L.; Xu, X.; Jiang, J.-H.; Shen, G.-L.; Yu, R.-Q. J. Am. Chem. Soc. 2009, 131, 2478–2480. (33) Liu, J.; Cao, Z.; Lu, Y. Chem. Rev. 2009, 109, 1948–1998. (34) Li, D.; Song, S.; Fan, C. Acc. Chem. Res. 2010, 43, 631–641. (35) Seelig, G.; Soloveichik, D.; Zhang, D. Y.; Winfree, E. Science 2006, 314, 1585–1588. (36) Johnson, S. M.; Grosshans, H.; Shingara, J.; Byrom, M.; Jarvis, R.; Cheng, A.; Labourier, E.; Reinert, K. L.; Brown, D.; Slack, F. J. Cell 2005, 120, 635–647. (37) Zhang, S.; Yan, Y.; Bi, S. Anal. Chem. 2009, 81, 8695–8701. (38) Lusi, E. A.; Passamano, M.; Guarascio, P.; Scarpa, A.; Schiavo, L. Anal. Chem. 2009, 81, 2819–2822. (39) Peng, Y.; Gao, Z. Anal. Chem. 2011, 83 (3), 820–827. (40) Cui, H.; Wang, W.; Duan, C.-F.; Dong, Y.-P.; Cuo, J.-Z. Chem. —Eur. J. 2007, 13, 6975–6984. (41) Nelson, P. T.; Baldwin, D. A.; Scearce, L. M.; Oberholtzer, J. C.; Tobias, J. W.; Mourelatos, Z. Nat. Methods 2004, 1, 155–161. (42) Fan, Y.; Chen, X.; Trigg, A. D.; Tung, C.-H.; Kong, J.; Gao, Z. J. Am. Chem. Soc. 2007, 129, 5437–5443.

3702

dx.doi.org/10.1021/ac200096b |Anal. Chem. 2011, 83, 3696–3702