Isothermal Nucleic Acid Amplification Strategy by Cyclic EnzymaticRepairing for Highly Sensitive MicroRNA DetectionDian-Ming Zhou, Wen-Fang Du, Qiang Xi, Jia Ge, and Jian-Hui Jiang*

State Key Laboratory of Chemo/Biosensing and Chemometrics, College of Chemistry and Chemical Engineering, Hunan University,Changsha 410082, P. R. China

*S Supporting Information

ABSTRACT: Technologies enabling highly sensitive andselective detection of microRNAs (miRNAs) are critical formiRNA discovery and clinical theranostics. Here we develop anovel isothermal nucleic acid amplification technology basedon cyclic enzymatic repairing and strand-displacementpolymerase extension for highly sensitive miRNA detection.The enzymatic repairing amplification (ERA) reaction isperformed via replicating DNA template using lesion basesby DNA polymerase and cleaving the DNA replicate at thelesions by repairing enzymes, uracil-DNA glycosylase, andendonuclease IV, to prime a next-round replication. Byutilizing the miRNA target as the primer, the ERA reactionis capable of producing a large number of reporter sequencesfrom the DNA template, which can then be coupled to a cyclic signal output reaction mediated by endonuclease IV. The ERAreaction can be configured as a single-step, close-tube, and real-time format, which enables highly sensitive and selective detectionof miRNA with excellent resistance to contaminants. The developed technology is demonstrated to give a detection limit of 0.1fM and show superb specificity in discriminating single-base mismatch. The results reveal that the ERA reaction may provide anew paradigm for efficient nucleic acid amplification and may hold the potential for miRNA expression profiling and relatedtheranostic applications.

MicroRNAs (miRNAs), a group of short (approximately19−25 nucleotides) and endogenous nonprotein-coding

RNA molecules, play crucial regulatory roles in gene expressionthat are critical to various biological processes. Aberrantexpression of miRNAs is closely implicated in various diseases,including cancers, diabetes, neurological disorders, cardiovas-cular, and autoimmune diseases.1 It is revealed that miRNAshas emerged as a new class of promising biomarkers for clinicaldiagnostics and targets for drug development and diseasetherapy.2 MiRNAs have several unique characteristics, such asshort lengths, sequence homology among family members,lability to degradation, and low abundance in total RNAsamples,3 which increases the difficulty in miRNA-related targetdiscovery and clinical theranostics. For the reasons, it is veryimportant to develop sensitive and selective technologies formiRNA detection.Detection of miRNA has been conventionally implemented

using Northern blotting technology4 and DNA microarrays5

without a preliminary amplification step. The unfavorably lowsensitivity of these techniques, however, precluded theirapplications in clinical theranostics. Recently, there is increasinginterest in the development of nucleic acid amplificationtechnologies to improve the sensitivity in miRNA analysis, suchas reverse transcription quantitative PCR,6 rolling circleamplification,7−9 exponential strand-displacement amplifica-tion,10 modified invader amplification,11 hairpin-mediated

quadratic enzymatic amplification,12 exponential amplificationreaction,13 and duplex-specific nuclease signal amplification.14,15

Most of these amplification reactions involve a key mechanismof DNA nicking, in which a phosphodiester bond is cleavedcyclically by a nickase such that 3′ end at the nick site can act asa primer for a DNA polymerase to initiate the product ofnumerous DNA replicates. However, it is reported that thecombination of nickase and DNA polymerase may lead tononspecific amplification even in the absence of DNAtemplates,16,17 possibly causing false positive signal in thedetection. Hence, the development of alternative mechanismsthat enable highly efficient polymerase-based amplification withlow nonspecific background remains a great challenge in thearea.Herein, we develop a novel isothermal nucleic acid

amplification technology based on cyclic enzymatic repairingand strand-displacement polymerase extension for highlysensitive miRNA detection. The enzymatic repairing amplifica-tion (ERA) reaction is primed specifically by the target miRNAand performed cyclically to amplify the designed DNA templateinto a large number of DNA replicates. The cyclic amplification

Received: May 19, 2014Accepted: June 20, 2014Published: June 20, 2014

Letter

pubs.acs.org/ac

© 2014 American Chemical Society 6763 dx.doi.org/10.1021/ac501857m | Anal. Chem. 2014, 86, 6763−6767

reaction comprises two steps: the replication of the DNAtemplate using lesion bases catalyzed by a DNA polymerase andthe scissoring at the lesions to create a new priming site by twoDNA repairing enzymes, uracil-DNA glycosylase (UDG) andendonuclease IV (Endo IV). Because nonspecific polymer-ization of nucleotides may incorporate too many lesions toform long DNA templates that can be efficiently amplified, thedeveloped ERA may have the potential to combat nonspecificbackground. Moreover, by using site-specific cleavage of afluorescence-quenched probe by Endo IV, we can directlyincorporate a real-time fluorescence-activation strategy forspecific and efficient detection of the amplified DNA replicates.This provides an additional advantage of further signalamplification over current nickase-based amplification reac-tions, which generally rely on double-strand fluorescencestaining13,18 or molecular beacon design19,20 for signal output.The analytical principle of the developed ERA technology for

miRNA detection is illustrated in Scheme 1. A DNA template isdesigned to comprise four regions and two lesion incorporatingsites: a target annealing region (T) that is complementary to agiven fragment at 3′ end of target miRNA, a primer-annealingregion (P1) that is used to hybridize with the primers for cyclicamplification, a reporter coding region (R) that is used togenerate a reporter sequence complementary to the fluo-rescence-quenched signal probe, and a primer-producing region(P2) that has an identical sequence as P1 and is used for primerproduction in cyclic amplification. In the assay, the miRNA isannealed on the DNA template in the T region, which can actas a primer for Bst DNA polymerase to initiate DNA extensionin the presence of four nucleotides dATP, dGTP, dCTP, anddUTP. (Note: dUTP is used instead of dTTP to incorporatelesions). The replication of the DNA template incorporates inthe DNA extension product two uracil (dU) nucleotides. Thesetwo lesion sites are excised by UDG via catalyzing thehydrolysis of the N-glycosylic bond joining the uracil base tothe deoxyribose,21,22 creating two abasic sites (AP site) that arethen cleaved by Endo IV at the phosphodiester bond upstreamto the abasic site.23 The repairing reaction, which generates ahydroxyl group at the 3′ terminus, will reprime the polymeraseextension from the lesion site, thereby generating another copyof the DNA template and displacing the first copy away fromthe template. The repetition of the polymerase extension andthe DNA repairing, therefore, results in a linear amplification ofthe DNA template into numerous copies of the primer-producing region P2 and the reporter coding region. The

displaced primers from P2 are then annealed on P1 of anotherDNA template, triggering a cycle of the polymerization-repairing amplification. This cyclic ERA reaction cannot onlygenerate a large number of copies of the primer for highlyefficient amplification but also produce a great quantity ofreporter sequences that are responsible for efficient signaloutput. To deliver an activated fluorescence signal, afluorescence-quenched signal probe24−26 is designed to have atetrahydrofuran abasic site mimic (TAP site) flanked in closeproximity by nucleotides modified with a fluorophore (FAM)and a quencher (Dabcyl). This signal probe is reported to be astable substrate for Endo IV cleavage.27 When annealed on thereporter sequence, the signal probe can be specifically andefficiently cleaved by Endo IV. The resulting two fragments,one carrying the fluorophore and the other bearing thequencher, are both too short to be stably annealed on thereporter sequence. The dissociation of these two fragmentsfrom the reporter sequence, therefore, activates the fluores-cence signal while enables signal amplification via the cyclicannealing and cleavage of the signal probe on the reportersequence. On the basis of this principle, the developed miRNAdetection technology can be implemented by a single-stepmixing of the miRNA sample and the enzyme system followedby a real-time monitoring of the fluorescence signal with noneed for additional steps for reagent inputs and reactions. Thissingle-step, closed-tube and real-time configuration can offerexcellent resistance to possible contaminants. Moreover,because of the combination of cyclic ERA amplification andEndo IV mediated signal amplification, the developedtechnology indeed provides a new paradigm for highly efficientnucleic acid amplification enabling very sensitive miRNAdetection. To demonstrate the feasibility of this technology,we choose miR-21 (Table S1 in the Supporting Information) asthe model system. MiR-21 has been reported to be overex-pressed in varying human cancers, including glioblastoma,cholangiocarcinoma, multiple myeloma cells, and breastcancer.28

A major advantage of the developed technology is theefficient inhibition for nonspecific amplification from template-or primer-independent DNA synthesis (Figure S1 in theSupporting Information). It was observed that, in the presenceof nickase Nt.BstNBI and DNA polymerase (Bst exo− or Ventexo−) as well as four nucleotides dATP, dGTP, dCTP, anddTTP, a great amount of DNA was synthesized at 60 °C evenin cases when no DNA was added as the template or primer.

Scheme 1. Illustration of ERA Strategy for miRNA Detection

Analytical Chemistry Letter

dx.doi.org/10.1021/ac501857m | Anal. Chem. 2014, 86, 6763−67676764

Such nonspecific DNA amplification is attributed to the DNApolymerase mediated ab initio synthesis29 and elongation ofDNA duplexes in which the short-sequence recognition site ofnickase will be randomly incorporated and used as seeds forfurther elongation.16,17 The repeated cycles of nickase mediateddigestion and polymerase based elongation thus allowexponentially amplification of nonspecific DNA under iso-thermal conditions. On the other hand, when DNA polymerasereacted with four nucleotides in the absence of nickase, nosubstantial DNA synthesis was obtained. This result impliedthat the cooperative reactions of nickase and DNA polymerasewere essential to the nonspecific DNA synthesis. In contrast,for the ERA reaction in which DNA polymerase, UDG, andEndo IV were incubated with four nucleotides dATP, dGTP,dCTP, and dUTP, nonspecific DNA was not detectable,suggesting the ability of the ERA technology to effectively avoidnonspecific amplification. Presumably, the ERA reaction canincorporate in the ab initio synthesis of DNA many dUnucleotides that can be excised as lesion sites by UDG. Thisprevents the formation of long double-stranded DNA that canbe used as templates for further amplification and hence avoidseffective amplification in the absence of DNA template.Figure 1 depicts the typical real-time fluorescence curves of

the ERA technology in the detection of miR-21. It was

observed that, in the presence of miR-21 (1 nM), thefluorescence intensity increases rapidly after ∼49 min andreaches a plateau after ∼90 min, displaying a typical sigmoidalresponse (curve a). In contrast, there was no substantialfluorescence activation in the control experiments where one ofthe reagents in the ERA reaction was not added. This resultverified the essential roles of the DNA template, the DNArepairing steps and DNA polymerase in the ERA reaction. Inanother control where dUTP was replaced by dTTP in thereaction mixture, there was also no appreciable fluorescenceincrease throughout the reaction, implying the incorporation ofdUTP lesions in DNA replication was necessary for the ERAreaction. In the experiments with blank sample (miR-21 targetnot added in the reaction) or nonhomologous target (miR-141), we obtained sigmoidal fluorescence intensity curves with

a steep increase after ∼131 min. The background responses areascribed to the primer-independent replication of the DNAtemplate, which is frequently involved in polymerase reaction.30

Because the background responses required much longer timeto reach the inflections than that for the positive sample, wecould define the point of inflection (POI), the timecorresponding to the maximum slope in the fluorescencecurve, for the quantification. Accordingly, we obtained a POIvalue of 49 min for 1 nM target and a POI value of 131 min forthe backgrounds. Such a big difference of the POI valuesrevealed that the ERA technique enabled effective differ-entiation of target miR-21 from nonhomologous targets. Inaddition to the real-time fluorescence monitoring of the ERAreaction, the steady-state fluorescence spectra were alsocollected after 90 min. The data confirmed that the ERAamplification delivered a high signal-to-background ratio fortarget miRNA and a nonhomologous target (Figure S2 in theSupporting Information). Gel electrophoresis analysis alsoverified the effective amplification of the reporter sequencesand the primers in the ERA reaction (Figure S3 in theSupporting Information).The ERA reaction enabled quantitative detection of target

miR-21 using the real-time fluorescence curve, as shown inFigure 2. It was observed that the real-time fluorescence curvesall displayed the sigmoidal shape, and the correspondinginflection regions appeared at shorter reaction time withincreasing concentrations of target miRNA in a wide

Figure 1. Real-time fluorescence curves of ERA reactions. (a) 1 nMmiR-21, 500 pM template, 100 nM signal probe, 0.05 U/μL Bst DNApolymerase, 0.08 U/μL UDG, 0.16 U/μL Endo IV, and 500 nMdNTPs (dATP, dGTP, dCTP, dUTP); (b) control with no DNAtemplate; (c) control with no UDG; (d) control with no Endo IV; (e)control with no DNA polymerase; (f) control with dUTP replaced bydTTP; (g) control with no miRNA target; (h) control with miR-21replaced by miR-141.

Figure 2. (A) Real-time fluorescence curves of ERA strategy for miR-21 detection. (B) Correlation of POI values to logarithmic miR-21concentrations. Error bars are standard deviations of four repetitiveexperiments.

Analytical Chemistry Letter

dx.doi.org/10.1021/ac501857m | Anal. Chem. 2014, 86, 6763−67676765

concentration range from 1 fM to 1 nM. A plot of the POIvalues versus the logarithmic concentrations of miR-21 showedlinear dependency in the miR-21 concentration range fromfrom 100 fM to 1 nM. The detection limit was calculated to be0.1 fM according to the POI value 3 times the standarddeviation over the blank response. Such a low detection limitimplied a superior sensitivity of the ERA strategy in miRNAdetection over existing methods.8,14,31,32 The improvedsensitivity might be attributed to the low background of theERA reaction as well as its high amplification efficiency and theadditional signal amplification mediated by Endo IV-basedcleavage. The relative standard deviations of POI values were4.9%, 3.2%, and 1.14%, respectively, for four repetitive assays ofmiR-21 of 100 pM, 1 pM, and 10 fM, implying excellentreproducibility of the ERA strategy.The ERA reaction not only provided improved detection

sensitivity but also afforded ideal specificity for target miRNAdetection, as shown in Figure 3. It was found that non-

homologous miRNA such as let-7a and miR-141 gave the samePOI value as the blank. Moreover, we observed that miRNAwith the single-base mismatch at the 3′ terminus also did notcause any interference, while single-base mismatch at other sitescaused slightly decreased POI values (<4 min). Such deviationsof POI values were much smaller than that (∼30 min) for theperfectly matched miRNA with a discrimination ratio >7. Thishigh specificity with mismatch discrimination ability wasderived from the miRNA-primed extension step in the ERAreaction, which was dominated by the perfect match at the 3′end and highly dependent upon the hybridization stabilitybetween the target and the template.To demonstrate the capability of the ERA technology for

miRNA detection in complex samples, miR-21 in total RNAextracts from five human cancer cell lines was analyzed usingthe developed strategy (Figure S4 in the SupportingInformation). These results showed that the ERA strategygave quantification data consistent with those obtained by RT-qPCR (The maximum relative deviation was 11%), implyingthe potential of the developed ERA technology for miRNAquantification in real complex samples. In addition, it wasrevealed that the miR-21 showed different expression levels inthese cancer cell lines with the human breast cancer cell lineMCF-7 having the highest expression of miR-21 in all the cell

lines, which was in good agreement with the data previouslyreported.33

In conclusion, we developed a novel isothermal nucleic acidamplification technology, enzymatic repairing amplification(ERA), based on cyclic enzymatic repairing and strand-displacement polymerase extension for highly sensitivemiRNA detection. Compared with commonly used nickasebased nucleic acid amplification technologies, the ERA reactionprovided comparable amplification efficiency with improvedresistance to nonspecific amplification independent of primersand templates. When used for miRNA detection, the ERAstrategy could be implemented in a single-step, closed-tube, andreal-time detection format, enabling convenient operation andresistance to contaminants. The developed strategy was shownto give a detection limit as low as 0.1 fM for miRNA detectionacross a wide dynamic range up to 1 nM. Moreover, it wasdemonstrated that this technique showed excellent specificitywith the capability of discriminating single-base mismatch andhad the potential to be used for miRNA detection in realcomplex samples. In virtue of these advantages, the proposedERA reaction indeed provided a new paradigm for efficientnucleic acid amplification and might hold the potential formiRNA expression profiling and related theranostic applica-tions.

■ ASSOCIATED CONTENT

*S Supporting InformationExperimental details and additional figures. This material isavailable free of charge via the Internet at http://pubs.acs.org.

■ AUTHOR INFORMATION

Corresponding Author*Phone: 86-731-88821916. Fax: 86-731-88821916. E-mail:jianhuijiang@hnu.edu.cn.

NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTS

This work was supported by NSFC (Grants 21025521,21221003, 21205034, 21035001, 21190041, 91317312) andthe National Key Basic Research Program (Grant2011CB911000).

■ REFERENCES(1) Arenz, C. Angew. Chem., Int. Ed. 2006, 45, 5048−5050.(2) Croce, C. M. Nat. Rev. Genet. 2009, 10, 704−714.(3) Baker, M. Nat. Methods 2010, 7, 687−692.(4) Lagos-Quintana, M.; Rauhut, R.; Lendeckel, W.; Tuschl, T.Science 2001, 294, 853−858.(5) Thomson, J. M.; Parker, J.; Perou, C. M.; Hammond, S. M. Nat.Methods 2004, 1, 47−53.(6) 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.(7) Jonstrup, S. P.; Koch, J.; Kjems, J. RNA 2006, 12, 1747−1752.(8) Cheng, Y.; Zhang, X.; Li, Z.; Jiao, X.; Wang, Y.; Zhang, Y. Angew.Chem., Int. Ed. 2009, 48, 3268−3272.(9) Deng, R.; Tang, L.; Tian, Q.; Wang, Y.; Lin, L.; Li, J. Angew.Chem., Int. Ed. 2014, 126, 2421−2425.(10) Shi, C.; Liu, Q.; Ma, C.; Zhong, W. Anal. Chem. 2014, 86, 336−339.

Figure 3. Specificity evaluation of ERA strategy for miRNA detection.Each miRNA has a concentration of 1 pM. ΔPOI is defined as thedifference of the POI values between the positive sample and theblank.

Analytical Chemistry Letter

dx.doi.org/10.1021/ac501857m | Anal. Chem. 2014, 86, 6763−67676766

(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) Duan, R.; Zuo, X.; Wang, S.; Quan, X.; Chen, D.; Chen, Z.;Jiang, L.; Fan, C.; Xia, F. J. Am. Chem. Soc. 2013, 135, 4604−4607.(13) Jia, H.; Li, Z.; Liu, C.; Cheng, Y. Angew. Chem., Int. Ed. 2010, 49,5498−5501.(14) Yin, B. C.; Liu, Y. Q.; Ye, B. C. J. Am. Chem. Soc. 2012, 134,5064−5067.(15) Xi, Q.; Zhou, D. M.; Kan, Y. Y.; Ge, J.; Wu, Z. K.; Yu, R. Q.;Jiang, J. H. Anal. Chem. 2014, 86, 1361−1365.(16) Liang, X.; Jensen, K.; Frank-Kamenetskii, M. D. Biochemistry2004, 43, 13459−13466.(17) Tan, E.; Erwin, B.; Dames, S.; Ferguson, T.; Buechel, M.; Irvine,B.; Voelkerding, K.; Niemz, A. Biochemistry 2008, 47, 9987−9999.(18) Van Ness, J.; Van Ness, L. K.; Galas, D. J. Proc. Natl. Acad. Sci.U.S.A. 2003, 100, 4504−4509.(19) Wang, G. L.; Zhang, C. Y. Anal. Chem. 2012, 84, 7037−7042.(20) Yin, B. C.; Liu, Y. Q.; Ye, B. C. Anal. Chem. 2013, 85, 11487−11493.(21) Krokan, H.; Standal, R.; Slupphaug, G. Biochem. J. 1997, 325, 1−16.(22) Xiang, Y.; Lu, Y. Anal. Chem. 2012, 84, 9981−9987.(23) Levin, J. D.; Johnson, A. W.; Demple, B. J. Biol. Chem. 1988,263, 8066−8071.(24) Liu, J. W.; Lu, Y. J. Am. Chem. Soc. 2007, 129, 9838−9839.(25) Ali, M. M.; Aguirre, S. D.; Lazim, H.; Li, Y. F. Angew. Chem., Int.Ed. 2011, 50, 3751−3754.(26) Huang, P. J.; Lin, J.; Cao, J.; Vazin, M.; Liu, J. Anal. Chem. 2014,86, 1816−1821.(27) Piepenburg, O.; Williams, C. H.; Stemple, D. L.; Armes, N. A.PLoS Biol. 2006, 4, e204.(28) Catuogno, S.; Esposito, C. L.; Quintavalle, C.; Cerchia, L.;Condorelli, G.; De Franciscis, V. Cancers 2011, 3, 1877−1898.(29) Ogata, N.; Miura, T. Nucleic Acids Res. 1998, 26, 4652−4656.(30) Ogata, N.; Miura, T. Biochem. J. 1997, 324, 667−671.(31) Hartig, J. S.; Grune, I.; Najafi-Shoushtari, S. H.; Famulok, M. J.Am. Chem. Soc. 2004, 126, 722−723.(32) Li, C.; Li, Z.; Jia, H.; Yan, J. Chem. Commun. 2011, 47, 2595−2597.(33) 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.

Analytical Chemistry Letter

dx.doi.org/10.1021/ac501857m | Anal. Chem. 2014, 86, 6763−67676767