Design of Molecular Beacons as Signaling Probesfor Adenosine Triphosphate Detection in CancerCells Based on Chemiluminescence ResonanceEnergy Transfer

Shusheng Zhang,* Yameng Yan, and Sai Bi*

Key Laboratory of Eco-chemical Engineering, Ministry of Education, College of Chemistry and Molecular Engineering,Qingdao University of Science and Technology, Qingdao 266042, P. R. China

In the present study, binary and triplex DNA molecularbeacons, as signaling probes based on a luminol-H2O2-horseradish peroxidase (HRP)-fluorescein chemilu-minescence resonance energy transfer (CRET) systemand structure-switching aptamers for highly sensitivedetection of small molecules, are developed usingadenosine triphosphate (ATP) as a model analyte todemonstrate the generality of the strategy. This CRETprocess occurs from donor luminol to acceptor fluo-rescein, which is oxidized by H2O2 and catalyzed byHRP. DNA aptamer for ATP is first attached on thesurface of magnetic nanoparicles (MNPs). The cDNAlinker has an extension that hybridizes with two otherDNAs (LumAuNP-DNA and F-DNA) or three otherDNAs (HRP-DNA, LumAuNP-DNA, and F-DNA) tofabricate CRET-BMBP-MNP or CRET-TMBP-MNP con-jugates that provide the CRET signals. Thus, in theabsence of ATP, when the MNPs are removed from thesolution, they also take with them the linker DNA andthe CRET signal probes, and no CRET signal can bedetected. However, when ATP is introduced, a com-petition for the ATP aptamer between ATP and thecDNA linker occurs. As a result, CRET-BMBP andCRET-TMBP are forced to dissociate from the MNPsurface based on the structure switching of the aptam-er. The CRET signals are proportional to the concen-tration of ATP. In order to accelerate the rate of theaptamer structure-switching process, an invader DNAis introduced into the proposed strategy. The presentCRET system provides a low detection limit of 1.1 ×10-7and 3.2 × 10-7 M for ATP detection by BMBPand TMBP, respectively, which also exhibits a goodselectivity for ATP detection. Sample assays of ATP inK562 leukemia cells and 4T1 breast cancer cellsconfirm the reliability and practicality of the protocol,which reveal a good prospect of this platform forbiological sample analysis.

Aptamers are single-stranded nucleic acids isolated fromrandom-sequence DNA or RNA libraries by an in vitro selection

process termed SELEX (systematic evolution of ligands byexponential enrichment).1,2 Aptamers are able to recognize abroad range of molecular targets including small molecules, ions,proteins, and whole cells.3-5 The versatility of target, the highaffinity, and specificity of aptamers, associated with the simplicityof in vitro selection, make aptamers attractive as molecularreceptors and sensing elements for bioanalytical applications.6-8

Adenosine triphosphate (ATP), well-known as energy currencyin cells, plays a fundamental role in cell metabolism which isinvolved in many metabolic processes.9 It has been found thatATP is an important substrate in living organisms and an indicatorfor cell viability and cell injury.10 The DNA aptamer for ATP wasfirst selected by Huizenga and Szostak with an equilibriumdissociation constant (Kd) of 6 ± 3 µM which appeared as a weaktarget-binding affinity compared with other aptamers, such asKd at the nanomolar level for a thrombin-binding aptamer.11,12

However, detection of the target molecules with low-affinityaptamers tends to be more challenging than that with high-affinityaptamers, which can be more evident to demonstrate the general-ity and effectivity of a strategy. Therefore, detection of ATP is ofgreat significance.

So far, various designs of aptamer-based sensors have beenadvanced for transducing aptamer-ATP interactions into electro-chemical,13,14 colorimetric,15 and fluororescent16,17 signals. How-

* Corresponding author. Tel.: +86-532-84022750. Fax: +86-532-84022750.E-mail: shushzhang@126.com (S.Z.); bisai11@126.com (S.B.).

(1) Tuerk, C.; Gold, L. Science 1990, 249, 505–510.(2) Ellington, A. D.; Szostak, J. W. Nature 1990, 346, 818–822.(3) Gold, L.; Polisky, B.; Uhlenbeck, O.; Yarus, M. Annu. Rev. Biochem. 1995,

64, 763–797.(4) Navani, N. K.; Li, Y. F. Curr. Opin. Chem. Biol. 2006, 10, 272–281.(5) Shangguan, D.; Li, Y.; Tang, Z.; Cao, Z.; Chen, H.; Mallikaratchy, P.; Sefah,

K.; Yang, C.; Tan, W. Pro. Natl. Acad. Sci. U.S.A. 2006, 103, 11838–11843.(6) Hansen, J. A.; Wang, J.; Kawde, A.-N.; Xiang, Y.; Gothelf, K. V.; Collins, G.

J. Am. Chem. Soc. 2006, 128, 2228–2229.(7) Cash, K. J.; Ricci, F.; Plaxco, K. W. J. Am. Chem. Soc. 2009, 131, 6955–

6957.(8) Freeman, R.; Sharon, E.; Tel-Vered, R.; Willner, I. J. Am. Chem. Soc. 2009,

131, 5028–5029.(9) Berg, J. M.; Tymoczko, J. L.; Stryer, L. In Biochemistry; WH Freeman: New

York, 2002.(10) Perez-Ruiz, T.; Martınez-Lozano, C.; Tomas, V.; Martın, J. Anal. Bioanal.

Chem. 2003, 377, 189–194.(11) Huizenga, D. E.; Szostak, J. W. Biochemistry 1995, 34, 656–665.(12) Deng, Q.; German, I.; Buchanan, D.; Kennedy, R. T. Anal. Chem. 2001,

73, 5415–5421.(13) Zuo, X.; Xiao, Y.; Plaxco, K. W. J. Am. Chem. Soc. 2009, 131, 6944–6945.(14) Zuo, X.; Song, S.; Zhang, J.; Pan, D.; Wang, L.; Fan, C. J. Am. Chem. Soc.

2007, 129, 1042–1043.

Anal. Chem. 2009, 81, 8695–8701

10.1021/ac901759g CCC: $40.75 2009 American Chemical Society 8695Analytical Chemistry, Vol. 81, No. 21, November 1, 2009Published on Web 09/29/2009

ever, while each strategy has distinct advantages, each alsopresents its own unique set of limitations. Despite a few groupswho have attempted to fabricate simple and convenient colori-metric sensing systems based on an aptamer probe, a highsensitivity could not been obtained. A fluorescent “signalingaptamer” possesses adequate transducing elements to generatephysically detectable signals from the recognition events, thusdesigning fluorescence-signaling aptamers have become the focalpoint of several recent studies. Because standard DNA and RNAdo not contain intrinsically fluorescent groups, aptamers must bemodified with external fluorophores to provide opportunities forfluorescence-based sensing. However, even slight modificationson aptamers may lead to significant loss of their affinity andspecificity, and fairly small fluorescence responses are usuallyobtained upon target binding; therefore, their detection sensitivityis relatively low.18-20 Recently, a few label-free aptamer biosensorshave been fabricated which use either photoactive polymer orDNA-intercalating dyes to report conformational change of aptam-ers upon recognition of given targets.21-24 These methods,however, suffer from intrinsic limitations of high background,nonselectivity against specific aptamer, and lack of multiplexdetection capability, therefore, limiting their applicability formultiplex or in situ detection. Consequently, the development ofa feasible detection method with high sensitivity, simplicity, andlow cost has become highly focused which can provide possiblyan alternative detection approach to the fluororescence-signalingaptamer system. Among them, a chemiluminescence (CL) detec-tion system has been one of the most attractive analytical toolswhich possesses characteristics of rapidity, precision, low cost,and simple manipulation.25,26 Yet, only a few aptamer-basedsensing platforms have been proposed so far for the CL detectionof small molecules.

Chemiluminescence resonance energy transfer (CRET) in-volves nonradiative (dipole-dipole) transfer of energy from achemiluminescent donor to a suitable acceptor molecule.27 Forthe luminol-H2O2-horseradish peroxidase (HRP)-fluorescein CLsystem, fluorescein could absorb part of the excited-stateluminol energy and re-emit it at longer wavelengths, whichresults in a CRET process from donor luminol (around 425nm) to acceptor fluorescein (around 510 nm).28 In contrast tofluorescence resonance energy transfer (FRET), CRET occursby the oxidation of a luminescent substrate without an excita-tion source. Inspired by this observation, we designed noveland sensitive binary and triplex DNA molecular beacons as

signaling probes (named BMBPs and TMBPs, respectively),which acheived the detection of celluar ATP for cancer cellsbased on CRET and a structure-switching aptamer.

EXPERIMENTAL SECTIONChemicals and Apparatus. All oligonucletides used in the

present study were synthesized and purified by SBS GenetechCo. Ltd. (China). Benzo[e]pyridoindole (BePI) was purchasedfrom Acros Organics (Geel, Belgium). Tri(2-carboxyethyl)phos-phine hydrochloride (TCEP, 98%) was purchased from Alfa Aesar(Ward Hill, MA). Streptavidin-HRP, luminol standard powder,1-ethyl-3-(3-dimethylaminopropyl)-carbodiimide hydrochloride(EDC), and N-hydroxysuccinimide (NHS) were ordered fromSigma-Aldrich. A luminol stock solution (1.0 × 10-2 M) wasprepared by dissolving it in 0.1 M NaOH solution and storingit in dark. HAuCl4 ·4H2O was ordered from Shanghai Reagent(Shanghai, China). A 1% (w/w) HAuCl4 stock solution wasprepared by dissolving 1 g of HAuCl4 in 100 mL of doublydistilled water, and it was stored at 4 °C for further use.Adenosine triphosphate (ATP), cytosine triphosphate (CTP),guanosine triphosphate (GTP), and uridine triphosphate (UTP)were obtained from Sigma, and their stock solutions (1.0 mM)were prepared by doubly distilled water. The resulting solutionwas further consecutively diluted with doubly distilled waterin order to obtain the proper solution used for CL detection.Doubly distilled water was used throughout the experiments.All the chemicals employed were of analytical reagent gradeand were used without further purification. Human serumsamples, provided by Qingdao Textile Hospital, were storedat 4 °C.

The CL measurements were performed using a BPCL ul-traweak luminescence analyzer (Institute of Biophysics AcademicSinica, Beijing, China), and the CL spectra were measured on aF-4500 fluorescence spectrophotometer (HITACHI, Japan). Trans-mission electron microscopy (TEM) and scanning electronmicroscopy (SEM) images were taken with a JEM-2000EX/ASID2and a JSM-6700F (HITACHI, Japan), respectively. UV-visiblespectra were taken with a Cary 50 UV-vis-NIR spectrophotometer(Varian). Magnetic nanoparticles (MNPs) modified with carboxylgroups (∼1.0 µm, 10 mg/mL) (COOH-MNPs) and amino groups(∼3.0 µm, 10 mg/mL) (NH2-MNPs) and a magnetic rack wereobtained from BaseLine Chrom Tech Research Centre (Tianjin,China).

Synthesis and Characterization of LumAuNPs. LumAuNPswere synthesized by following the method of reduction oftetrachloroauric acid with luminol which has been carried out byCui et al.29 Briefly, 100 mL of 0.02% (w/w) HAuCl4 solution wasboiled with vigorous stirring, and then, various amounts of 1.0× 10-2 M luminol stock solution (1.4, 1.8, 2.2, and 2.6 mL) wereadded to the boiling solution rapidly. The solution wasmaintained at the boiling point for 30 min. The color of thesolution turned from yellow to black to purple before a wine-red color was reached, indicating the formation of LumAuNPs.After cooling down the resulting colloidal suspension to roomtemperature with stirring, the colloidal LumAuNPs wereobtained. A dialysis operation was carried out to separateexcess luminol from the AuNPs.

(15) Wang, J.; Wang, L.; Liu, X.; Liang, Z.; Song, S.; Li, W.; Li, G.; Fan, C. Adv.Mater. 2007, 19, 3943–3946.

(16) Nutiu, R.; Li, Y. J. Am. Chem. Soc. 2003, 125, 4771–4778.(17) Nutiu, R.; Li, Y. Chem.sEur. J. 2004, 10, 1868–1876.(18) Beyer, S.; Dittmer, W. U.; Simmel, F. C. J. Biomed. Nanotechnol. 2005, 1,

96–101.(19) Jhaveri, S.; Rajendran, M.; Ellington, A. D. Nat. Biotechnol. 2000, 18, 1293–

1297.(20) Nutiu, R.; Li, Y. F. Angew. Chem., Int. Ed. 2005, 44, 1061–1065.(21) Li, N.; Ho, C.-M. J. Am. Chem. Soc. 2008, 130, 2380–2381.(22) Ho, H. A.; Leclerc, M. J. Am. Chem. Soc. 2004, 126, 1384–1387.(23) Stojanovic, M. N.; Landry, D. W. J. Am. Chem. Soc. 2002, 124, 9678–9679.(24) Jiang, Y. X.; Fang, X. H.; Bai, C. L. Anal. Chem. 2004, 76, 5230–5235.(25) Zhang, S.; Zhong, H.; Ding, C. Anal. Chem. 2008, 80, 7206–7212.(26) Bi, S.; Yan, Y.; Yang, X.; Zhang, S. Chem.sEur. J. 2009, 15, 4704–4709.(27) Huang, X.; Li, L.; Qian, H.; Dong, C.; Ren, J. Angew. Chem., Int. Ed. 2006,

45, 5140–5143.(28) Dıaz, A. N.; Garcıa, J. A. G.; Lovillo, J. J. Biolumin. Chemilumin. 1997, 12,

199–205.(29) Cui, H.; Wang, W.; Duan, C.-F.; Dong, Y.-P.; Guo, J. Z. Chem.sEur. J. 2007,

13, 6975–6984.

8696 Analytical Chemistry, Vol. 81, No. 21, November 1, 2009

The LumAuNPs, synthesized by different amounts of luminolsolution, were characterized with TEM (see Figure S1 in theSupporting Information), and UV-visible and fluorescence spectrawere also recorded (see Figure S2 in the Supporting Information).Gold colloid prepared with 1.80 mL of luminol was selected forthe modification procedures.

Preparation of LumAuNPs-Labeled Oligonucleotides. TheLumAuNP-oligonucleotide conjugates were synthesized by adding∼300 µL of the prepared aqueous LumAuNP solution (∼10 nM)to 100 µL of 100 µM thiol-modified DNA. Thiol-modified DNAwas activated with TCEP (10 mM) for 1 h before being attachedto LumAuNPs. After shaking gently for 16 h at room temperature,the LumAuNP-DNA conjugates were “aged” in the solution (0.3M NaCl, 10 mM Tris-acetate, pH 8.2) for another 48 h. Excessreagents were removed by centrifuging at 15 000 rpm for 30 min.Following removal of the supernatant, the resulting red oilyprecipitate was washed with 500 µL of 0.1 M pH 7.0 phosphatebuffer containing 0.1 M NaCl, recentrifuged, and then redispersedin 500 µL of 0.1 M pH 7.0 phosphate buffer containing 0.3 M NaCl.

Preparation of Aptamer-MNPs. The TEM image of thecarboxylated MNPs with an average diameter of 1.0 µm is shownin Figure S3 (in the Supporting Information). The aptamer-MNPswere prepared according to the reference with a slight modifica-tion. Briefly, a 30 µL suspension of carboxylated MNPs was placedin a 1.5 mL Eppendorf tube (EP tube) and separated from thesolution on a magnetic rack. After washing three times with 200µL of 0.01 M PBS buffer (pH 7.4), a 200 µL aliquot of 0.1 Mimidazol-HCl buffer containing 0.2 M EDC was added to the EPtube, and the mixture was incubated at 37 °C for 40 min to activatethe carboxylate groups on the MNPs, followed by washing threetimes with 200 µL of 0.01 M PBS buffer. Then, 100 µL of a 100µM amino group modified aptamer at a T9 spacer end was addedto the above freshly activated MNPs and incubated at 37 °Covernight. Finally, the resulting aptamer-MNPs were washedwith 200 µL of 0.01 M PBS buffer three times, resuspended in200 µL of PBS buffer, and stored at 4 °C for further use.

Fabrication of CRET-BMBP-MNPs. Stock solutions of threeDNA components, linker-DNA, LumAuNP-DNA, and F-DNA, wereprepared at a concentration of 1.0 mM in PBS buffer (pH 7.4).First, 100 µL of 100 µM linker-DNA was incubated with theprepared aptamer-MNPs and allowed to react in 0.1 M PBS bufferfor 60 min at 37 °C. After washing with 200 µL of 0.01 M PBSbuffer three times, 400 µL of 25 µM LumAuNP-DNA and 100 µLof 100 µM F-DNA in 0.01 M PBS buffer were added to the DNA-MNPs conjugates simultaneously; for 1 h, the samples were placedat 37 °C to perform the duplex hybridization with linker-DNA.The resulting BMBP-MNP conjugates were redispersed in 0.5 mLof 0.1 M pH 7.0 phosphate buffer containing 0.3 M NaCl, and theconjugates were stored at 4 °C for future use.

Fabrication of CRET-TMBP-MNPs. Stock solutions of threeDNA components, biotin-DNA with a linker, LumAuNP-DNA, andF-DNA, were prepared at a concentration of 1.0 mM in PBS buffer(pH 7.4). First, 100 µL of 100 µM biotin-DNA was incubated withthe prepared aptamer-MNPs and allowed to react in 0.01 M PBSbuffer for 60 min at 37 °C. To attach HRP to biotin-DNA to obtainan HRP labeled DNA strand (HRP-DNA), 100 µL of 100 µMstreptavidin-HRP was added and incubated for 40 min at 37 °C,followed by washing three times with 400 µL of 0.01 M PBS at

room temperature. Then, 400 µL of 25 µM LumAuNP-DNA and100 µL of 100 µM F-DNA in 0.01 M PBS buffer were added.Subsequently, 6 µL of 10 mM DNA binding molecules BePI wereadded. Incubation with the DNA binders was allowed to proceedfor 1 h at 37 °C. The resultant TMBP-MNPs were redispersed in0.5 mL of 0.1 M pH 7.0 phosphate buffer containing 0.3 M NaCl,and the conjugates were stored at 4 °C for future use.

Note that, for both the BMBPs and CMBPs preparations, theobtained MNPs conjugates were washed three times with PBSbuffer solution on a magnetic rack after each hybridizationprocedure.

Preparation of ATP Extracts from Cancer Cells. The K562leukemia cells and 4T1 breast cancer cells were separatelycultured in cell flasks according to the instructions from theAmerican Type Culture Collection. The cell line was grown to90% confluence in RPMI 1640 mCEedium supplemented with 10%fetal bovine serum (FBS) and 100 IU/mL penicillin-streptomycinat 37 °C, and the cells were harvested by trypsinization. The celldensity was determined by a hemocytometer prior to eachexperiment. Then, a suspension of 2.14 × 106 cells (3.0 mL) forK562 cells and 8.67 × 105 cells (3.0 mL) for 4T1 cells, dispersedin RPMI cell media buffer, was centrifuged at 1000 rpm for 5min and washed with phosphate-buffered saline (18.6 mMphosphate, 4.2 mM KCl, and 154.0 mM NaCl, pH 7.4) five timesand resuspended in 0.25 mL of deionized water. Finally, thecells were disrupted by sonication for 20 min at 0 °C. To removethe homogenate of cell debris, the lysate was centrifuged at18 000 rpm for 20 min at 4 °C. For comparison, ATP levelswere assayed with a modified HPLC method as described inthe Supporting Information.

ATP Analysis. ATP, both in the standard form and thatreleased from cells, was measured using the BMBP-MNPs andTMBP-MNPs. Various samples (200 µL) at a specific concentrationwere added and kept for 20 min to make the aptamer change itsstructure to bind ATP. After magnetic separation, 100 µL of thesupernatant was transferred to the quartz cuvette containing 100µL of 1.0 µM HRP, which is not needed for CRET-TMBPs. TheCL reaction was triggered by injecting 200 µL of 7.5 × 10-3 Mluminol with a syringe through a septum after the CL analyzerbegan to record at 10 s. The kinetics of CL signals between 0and 30 s was recorded, and the peak heights of the emissioncurves were measured by means of a photon counting unit.

RESULTS AND DISCUSSIONDesign of Molecular Beacons as Signaling Probes. Based

on the luminol-H2O2-HRP-fluorescein CRET process, the con-struction of the BMBPs and TMBPs in this study are shownas Scheme 1A,B. For the CRET-BMBPs (Scheme 1A), the CLdonor, LumAuNP, is labeled at 5′-end thiol-modified DNA (Lu-mAuNP-DNA), and the other probe is labeled with a fluoresceinmoiety at its 3′ end (F-DNA). In the absence of the linker DNA,the oligonucleotide strands are randomly dispersed in solution.However, in the presence of the linker DNA, a hybridization ofboth probes to the linker DNA is formed, which brings luminolmolecules in close proximity to fluorescein, and the enhanced CLof luminol-H2O2-HRP-fluorescein system is observed becauseCRET occurs from luminol donor to fluorescein acceptor. Forthe CRET-TMBPs (Scheme 1B), LumAuNP, fluorescein, andHRP are labeled to each end of three oligonucleotide strands,

8697Analytical Chemistry, Vol. 81, No. 21, November 1, 2009

which are further formed into triplex DNA structures by BePI, astrong triplex binder. Thus, a CRET process could occur fromluminol to fluorescein catalyzed by HRP after injecting H2O2 intothe triplex-structure probe system.

Furthermore, the reaction of the fluorescein enhanced luminol-H2O2-HRP CRET system was extensively studied. As shown inFigure 1A, the CL of the luminol-H2O2-HRP system and fluo-rescein could be found around the wavelengths of 425 and 510nm, respectively. Once fluorescein is introduced into theluminol-H2O2-HRP system, there are two maxima of emission,one (luminol) around 425 nm and another (fluorescein) around510 nm.

In addition, with the increasing concentration of fluorescein,the total emission between 380 and 600 nm is enhanced and theintensity at 510 nm increases while that at 425 nm decreases (seethe Supporting Information). These results could suggest that aCRET process occurs from donor luminol to acceptor fluorescein;that is, fluorescein absorbed part of the excited-state luminolenergy and re-emitted it at longer wavelengths. The mechanismis summerized in Figure 1B.

A control experiment was performed to support the CRETmechanism by employing complementary, two bases mutated anda non cDNA linker to hybridize with LumAuNP-DNA and F-DNA(see the Supporting Information). As shown in Figure S7 (in the

Supporting Information), the CRET signals could be obtained onlywhen F-DNA and LumAuNP-DNA hybridized with the cDNAlinker but cannot be observed for two bases mismatched and anon cDNA linker. The results strongly suggested that theobserved CRET signals occurred from donor luminol to acceptorfluorescein when LumAuNP-DNA and F-DNA hybridized with thecDNA linker which brought these two probes in close proximityto each other and resulted in an efficient CRET. In this experi-ment, magnetic Au@Fe3O4 was synthesized to overcome thehigh background of CRET-BMBPs based on the differentelectrostatic propensities of single- and double-stranded oligo-nucleotide (dsDNA and ssDNA) adsorption on gold nanopar-ticles.30 More importantly, the CRET-BMBPs in this study withhigh selectivity could be used to identify DNA sequence fortarget DNA hybridizaiton.

Employing Fabricated BMBPs and TMBPs for ATP De-tection. On the basis of the above constructed CRET-BMBP andCRET-TMBP platform, a highly sensitive detecion of ATP wasdeveloped through the conformational change of aptamers froma DNA/DNA duplex to a DNA/target complex upon the introduc-tion of target molecule, ATP, which is accompanied with a changeof CL intensity. The DNA-based ATP aptamer binds two ATPmolecules in a noncanonical, stable, helix composed of G:G and

(30) Li, H.; Rothberg, L. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 14036–14039.

Scheme 1. Schematics of CRET-BMBPs (A) and CRET-TMBPs (B) and Their Application in the Detection of ATP byCRET-BMBPs (C) and CRET-TMBPs (D)a

a The principle is described in the text for details.

8698 Analytical Chemistry, Vol. 81, No. 21, November 1, 2009

G:A base pairs flanked by short canonical helices.31 To realizethe full potential of DNA aptamers for applications such asmultianalyte biosensing, metabolite profiling, reporting enzymaticactivity, or affinity capture of specific analytes, it is necessary toimmobilize the aptamer on a sutiable surface.32 In the presentstudy (Scheme 1C,D), amino groups modified an aptamer probe,which specifically recognize ATP, by first attaching onto thesurface of carboxylated MNPs via EDC/NHS activation. CRET-BMBPs and CRET-TMBPs, whose linker DNA are designed tohave 12-mer to hybridize with aptamer sequences, are immobilizedon the MNP surface by aq hybridization reaction. The introductionof ATP triggered structure switching of the aptamer, and then,CRET-BMBPs or CRET-TMBPs are forced to dissociate from theMNP surface, resulting in an increase in the CL intensity. Thelight intensity of the CL produced in this reaction is proportionalto the amount of ATP present.

Disassembly Aided by Invasive DNA. Although 37 °C waschosen as the optimal incubation temperature for the ATP-aptamerbinding, the rate of response for the immobilized aptamer on the

solid surface was significantly slower than in solution, consistentwith much slower diffusion of the ATP toward the aptamer.

In order to circumvent this problem, a 12-mer DNA (calledinvader), which was complementary to the linker DNA, was addedto help release the CRET probes and accelerate the rate of theaptamer structure-switching process. For the TMBPs, it wasobserved that the structure-switching process was completed in∼10 min in the presence of ATP and 100 µM of the 12-mer invader(inva-1) (Figure. 2A). However, the background, in the absenceof ATP, was also increased, which can be ascribed to the invaderinvading not only the released linker DNA of CRET probes butalso the unreleased ones to a certain extent. To overcome thebackground interference, a series of shorter invader DNAs witha reduced number of base pairs on one end of the linker DNA ofthe MB probes were investigated (named inva-2, inva-3, and inva-4, for the sequences see Table S1 in the Supporting Information).

It can be assumed that the shorter invader should be morespecific toward the released linker DNA of CRET probes thanthe unreleased substrates. As shown in Figure 2B, the backgroundreleased in the absence of ATP was overcome by employing ashorter invader, although only a slight sacrifice of the release rate

(31) Lin, C. H.; Patel, D. J. Chem. Biol. 1997, 4, 817–832.(32) Rupcich, N.; Nutiu, R.; Li, Y.; Brennan, J. D. Anal. Chem. 2005, 77, 4300–

4307.

Figure 1. (A) CL spectra of fluorescein (green), luminol-H2O2-HRP (blue), and luminol-H2O2-HRP-fluorescein (red), respectively. (B) The CRETmechanism of the fluorescein enhanced luminol-H2O2-HRP system. Briefly, the fluorescein radical (F•-) increases the formation of the luminolradical (L•-). After the formation of luminol endoperoxide (LO2

2-), an energy transfer from LO22- to F•- forms a fluorescein endoperoxide (FO2

2-)which liberates oxygen and emits luminescence at the same time.

Figure 2. (A) Effect of inva-1 on the CRET-TMBPs at 37 °C. (B) Comparison of CRET-TMBP disassembly kinetics in the presence of differentinvaders.

8699Analytical Chemistry, Vol. 81, No. 21, November 1, 2009

of the CRET probes was in the presence of ATP. The illustrationfor this improved strategy employing invader is shown in SchemeS2 in the Supporting Information. A similar tendency was obtainedfor the BMBPs, and the results were not shown here. Consideringboth the rate of release and the level of background, inva-3 wasselected as trade-off for the subsequent analytical work.

Sensitivity and Selectivity of ATP Analysis by CRET-BMBPs and CRET-TMBPs. Incubation temperature and reac-tion time for the ATP and aptamers are the most two importantparameters to optimize the analysis system. After optimizing thesetwo factors (see the Supporting Information), reactions for 20 minat 37 °C were selected as a trade-off for the subsequent analyticalwork, and an invader with 8-mer complementary to the linker DNAwas added to help release the CRET probes and accelerate therate of the aptamer structure-switching process considering boththe rate of release and the level of background.

The introduction of ATP at different concentrations induceddifferent increases in CL signals associated with the amounts ofreleased CRET probes. Series samples of ATP were determinedby the present protocols, and the CL peak height was utilized toevaluate the CL response to ATP. As shown in Figure 3A, for theCRET-BMBPs, the CL signal increased with the increase of theconcentration of ATP, and the CL intensity revealed a linearrelationship with the concentration of ATP ranging from 1.0 × 10-6

to 1.0 × 10-5 M. The regression equation could be expressedas I ) 56.0273C + 17.4505 (I represents the peak height ofCL; C represents the concentration of ATP, µM; n ) 7, R )0.9979). As shown in Figure 3B, for the CRET-TMBPs, the CLintensity revealed a linear relationship with the concentration ofATP ranging from 1.0 × 10-6 to 1.0 × 10-5 M. The regressionequation could be expressed as I ) 44.3311C + 1.9625 (Irepresents the peak height of CL; C represents the concentra-tion of ATP, µM; n ) 7, R ) 0.9963). A detection limit of 1.1 ×10-7 and 3.2 × 10-7 M ATP can be estimated using 3σ for CRET-BMBPs and CRET-TMBPs (σ ) Sb/m; Sb, standard deviationof blank sample, Sb ) 2.05 and 4.73 (n ) 11) for CRET-BMBPsand CRET-TMBPs in this experiment; m, the slope of thecalibration curve). The standard curve was obtained by averag-ing three series of samples with the interassay variation of 6.2%for CRET-BMBPs and 8.8% for CRET-TMBPs.

From the calibration curves of Figure 3A,B, an apparent Kd,which was obtained by the target concentration that induceda half-maximal CL intensity change, was estimated as 20 µMfor ATP, which is only 2-fold to that reported for the original

ATP aptamer.11,12 The affinity maintenance can be explained asfollows: instead of modifying the aptamer, such as fluorophoreor electrochemical labels, the aptamer used in this study justmodified an amino group at the end of the spacer in order toimmoblize the aptamer on the MNP surface. Thus, there was onlya slight effect on the aptamer affinity and specificity, and theaptamer still retained its high activity to binding with the targets.

CL detection is one of the most sensitive methods for biologicalanalysis, and the sensitivities were higher than other techniquesfor the detection of ATP (Table 1); however, ATP, chosen as themodel in this study, is considered as a target molecule with low-affinity aptamers. Considering the low-affinity of the ATP-bindingaptamer, the relatively weak target-aptamer interaction does notcompete favorably with the aptamer-DNA interaction. Therefore,higher target concentrations are required to form the ATP-aptamercomplex. We hypothesize that, for the high-affinity aptamer (e.g.,the thrombin-binding aptamer), the strong binding interactionbetween the aptamer and the target could compete favorably withthe DNA-aptamer duplex hybridization interaction, which couldfairly demonstrate the generality of this strategy.

To assess the specificity of the method for the detection ofATP, experiments were conducted on CTP, UTP, and GTP.Different CL signals of the proposed system for the detection of60 µM ATP, CTP, UTP, or GTP were under the same experimentalconditions for CRET-BMBPs and CRET-TMBPs, respectively.

(33) Liu, B. F.; Ozaki, M.; Hisamoto, H.; Luo, Q.; Utsumi, Y.; Hattori, T.; Terabe,S. Anal. Chem. 2005, 77, 573–578.

(34) Tang, Z.; Mallikaratchy, P.; Yang, R.; Kim, Y.; Zhu, Z.; Wang, H.; Tan, W.J. Am. Chem. Soc. 2008, 130, 11268–11269.

(35) Huang, Y. F.; Chang, H. T. Anal. Chem. 2007, 79, 4852–4859.

Figure 3. Quantification of ATP concentration by measuring the CL intensity after the addition of ATP to CRET-BMBPs-MNPs (A) and CRET-TMBPs-MNPs (B).

Table 1. Comparison of the Detection of ATP Based ona Structure-Switching Aptamer

target detection limit (M) detection methoda reference

ATP 1.0 × 10-5 fluorescent detection 16ATP 1.0 × 10-5 fluorescent detection 32ATP 2.0 × 10-7 bioluminescence detection 33ATP 1.0 × 10-5 fluorescent detection 21ATP 5.0 × 10-4 fluorescent detection 34ATP 2.1 × 10-6 SAL-DI 35ATP 1.1 × 10-7 CRET-BMBPs bATP 3.2 × 10-7 CRET-TMBPs b

a SAL-DI, surface-assisted laser desorption/ionization; CRET, chemi-luminescence resonance energy transfer. b This work.

8700 Analytical Chemistry, Vol. 81, No. 21, November 1, 2009

On the basis of structure-switching signaling aptamers for thedetection of small molecules, CTP, UTP, or GTP did not induceany significant changes in the CL signal as compared to that ofATP for both CRET-BMBPs and CRET-TMBPs, which suggeststhat the proposed method has a high degree of specificity for thedetection of ATP and could become a general method for anyaptamer of interest.

Determination of ATP in a Cultured Cell Extract. To testthe validation of the platform for real-world samples, analysis ofcellular ATP from K562 leukemia cell and 4T1 breast cancer cellextract were implemented. The cell lysates were treated withdeproteination by filtration using cutoff membranes in order tominimize protein interferences. After adding the deproteinized celllysates into the CRET-BMBPs and CRET-TMBPs, the mixtureswere subjected to incubation for 20 min to ensure the full reactionbetween the aptamer and ATP. The results are listed in Table 2.The concentration of ATP in the K562 cell lysate is 3.7 (±0.3) mM(n ) 3) and 2.4 (±0.3) mM for 4T1 cells. For comparison, HPLCwas conducted for the analysis of the same sample (the detailsfor HPLC detection are described in the Supporting Information).The comparative results from these two methods were shown inFigure S10 in the Supporting Information; relatively good cor-relations were obtained as R ) 0.9997 and R ) 0.9996 for CRET-BMBPs and CRET-TMBPs, respectively.

CONCLUSIONIn summary, a novel and sensitive strategy to convert the

aptamer-target recognition event into CL signals employingBMBPs and TMBPs as signaling probes based on CRET isproposed in this study. This strategy has several significantadvantages. First, aptamers can be easily selected for diverseranges of biological targets with both high binding affinity andspecificity and can be created synthetically. In this study, thereis no need to make modification to the original aptmer, except anamino-group is modified at the spacer end in order to attach theaptamer to the MNPs which has little effect on the affinity andspecificity of the original aptamer. Second, the CRET-BMBPs andCRET-TMBPs are fabricated through partially hybridization with

the aptamer which was immobilized on the MNPs, preventing thedissociation of probe during application and making CRET-BMBPsand CRET-TMBPs robust molecular probes. After introductionof the target molecule, the signal transduction elements are forcedto dissolve into the solution, which is important for signalingtransduction in probe construction. Furthermore, an invader isintroduced to accelerate the rate of the aptamer structure-switching process. Third, in the control experiment to supportthe CRET mechanism, Au@Fe3O4 is used on the basis of thedifferent adsorption propensities of ssDNA and dsDNA on agold nanoparticle surface, considering the high backgroundand the separation procedure, enabling faster response andlower background, which provides an alternative approach forthe detection of DNA rival over the fluorescence binary probe.Finally, this strategy has been successfully employed in cancercellular ATP detection and can be of promise to be generallyused not only for different types of targets (biological cofactors,metabolites and proteins) but also for different types ofaptamers (DNA and RNA). In conclusion, with its simplicity,sensitivity, and specificity, this strategy as a molecular toolholds great promise in high-thoughput drug screening, in vitrodiagnostics, and intracellular studies.

ACKNOWLEDGMENTThe work was supported by the Excellent Young Scientists

Foundation of Shandong Province (JQ200805), the UniversityDoctoral Foundation of the Ministry of Education (200804260001),and the National Basic Research Program of China(2010CB732404).

SUPPORTING INFORMATION AVAILABLEAdditional information as noted in text. This material is

available free of charge via the Internet at http://pubs.acs.org.

Received for review August 5, 2009. Accepted September17, 2009.

AC901759G

Table 2. Comparisons of the CRET-Based Method with the HPLC Method for the Detection of ATP in K562 Cells and4T1 Cellsa

sampleb CRET-BMBPs-based method (mM) RSD (%) CRET-TMBPs-based method (mM) RSD (%) HPLC method (mM) RSD (%)

1 3.74 3.8 3.60 5.6 3.57 4.12 3.95 5.1 3.85 3.8 3.78 3.93 3.89 4.2 3.75 4.4 3.68 4.54 2.21 3.1 2.31 3.5 2.14 3.65 2.53 3.7 2.59 5.1 2.45 4.46 2.67 4.6 2.68 4.3 2.55 4.7

a Samples 1-3 were from K562 cells, and 4-6 were from 4T1 cells. b Each sample was analyzed in triplicate, and the results are the averagevalues.

8701Analytical Chemistry, Vol. 81, No. 21, November 1, 2009