Development of an Aptamer Beacon for Detectionof Interferon-Gamma

Nazgul Tuleuova,†,‡ Caroline N. Jones,† Jun Yan,† Erlan Ramanculov,‡ Yohei Yokobayashi,† andAlexander Revzin*,†

Department of Biomedical Engineering, University of California, Davis, California, and National Center forBiotechnology, Astana, Kazakhstan

Traditional antibody-based affinity sensing strategies employmultiple reagents and washing steps and are unsuitable forreal-time detection of analyte binding. Aptamers, on theother hand, may be designed to monitor binding eventsdirectly, in real-time, without the need for secondary labels.The goal of the present study was to design an aptamerbeacon for fluorescence resonance energy transfer (FRET)-based detection of interferon-gamma (IFN-γ)san importantinflammatory cytokine. Variants of DNA aptamer modifiedwith biotin moieties and spacers were immobilized onavidin-coated surfaces and characterized by surface plasmonresonance (SPR). The SPR studies showed that immobiliza-tion of aptamer via the 3′ end resulted in the best bindingIFN-γ (Kd ) 3.44 nM). This optimal aptamer variant wasthen used to construct a beacon by hybridizing fluoro-phore-labeled aptamer with an antisense oligonucleotidestrand carrying a quencher. SPR studies revealed thatIFN-γ binding with an aptamer beacon occurred within15 min of analyte introductionssuggesting dynamicreplacement of the quencher-complementary strand byIFN-γ molecules. To further highlight biosensing applica-tions, aptamer beacon molecules were immobilizedinside microfluidic channels and challenged with varyingconcentration of analyte. Fluorescence microscopy re-vealed low fluorescence in the absence of analyte andhigh fluorescence after introduction of IFN-γ. Impor-tantly, unlike traditional antibody-based immunoassays,the signal was observed directly upon binding of analytewithout the need for multiple washing steps. The surfaceimmobilized aptamer beacon had a linear range from 5to 100 nM and a lower limit of detection of 5 nM IFN-γ.In conclusion, we designed a FRET-based aptamerbeacon for monitoring of an inflammatory cytokinesIFN-γ. In the future, this biosensing strategy will be employedto monitor dynamics of cytokine production by theimmune cells.

Interferon-gamma (IFN-γ) is an important inflammatory cy-tokines secreted by immune cells in response to various patho-

gens.1 The levels of this protein provide diagnostic informationabout various infectious diseases and the ability of the body tomount an immune response. For example, in HIV infectedpatients, vigorous production of IFN-γ by T-helper (Th1) andcytotoxic T-lymphocytes correlates with low viremia and slowprogression of the disease.2,3 Traditionally, secreted cytokinessuch as IFN-γ are detected using antibody-based sandwichimmunoassays. While robust and well-established, these traditionalstrategies are time-consuming, require multiple washing steps,and provide no information about dynamics of cytokine production.

Aptamer-based affinity sensing strategies are emerging asviable alternatives to antibody-based immunoassays.4 Aptamersare single-stranded DNA or RNA oligonucleotides selected in vitroto bind target analytes with high specificity and affinity.5 Becauseaptamers are short nucleic acid molecules they are more robustthan antibodies so that aptamer-based biosensors can be regener-ated and used multiple times. Even more importantly, thesimplicity and robustness of aptamers makes them particularlyamenable to chemical modification and inclusion of surfacebinding or sensing moieties.6-9 Several strategies of transformingaptamer-analyte interactions into electrochemical, mechanical,piezoelectric, or fluorescent signals have been reported.10-15

Among these methods, fluorescence-based signal transduction isquite powerful because such strategies as fluorescence resonanceenergy transfer (FRET) may be utilized to convert aptamers into

* To whom correspondence should be addressed. Mailing address: Depart-ment of Biomedical Engineering University of California, Davis, 451 East HealthSciences Drive No. 2519, Davis, CA, 95616. E-mail: arevzin@ucdavis.edu. Phone:530-752-2383. Fax: 530-754-5739.

† University of California, Davis.‡ National Center for Biotechnology.

(1) Boehm, U.; Klamp, T.; Groot, M.; Howard, J. C. Annu. Rev. Immunol. 1997,15, 749–795.

(2) Pantaleo, G.; Koup, R. A. Nat. Med. 2004, 10, 806–810.(3) Romagnani, S. Clin. Immunol. Immunopath. 1996, 80 (3), 225–235.(4) Jayasena, S. D. Clin. Chem. 1999, 45, 1628–1650.(5) Ellington, A. D.; Szostak, J. W. Nature 1990, 346, 818–822.(6) Balamurugan, S.; Obubuafo, A.; Soper, S. A.; Spivak, D. A. Anal. Bioanal.

Chem. 2008, 390, 1009–1021.(7) Kirby, R.; Cho, E. J.; Gehrke, B.; Bayer, T.; Park, Y. S.; Neikirk, D. P.;

McDevitt, J. T.; Ellington, A. D. Anal. Chem. 2004, 76, 4066–4075.(8) Nutiu, R.; Li, Y. F. Methods 2005, 37, 16–25.(9) Luzi, E.; Minunni, M.; Tombelli, S.; Mascini, M. Trac-Trends Anal. Chem.

2003, 22, 810–818.(10) Lu, Y.; Zhu, N; Yu, P.; Mao, L. Anal. 2008, 133, 1256–1260.(11) Ikanovic, M.; Rudzinski, W. E.; Bruno, J. G.; Allman, A.; Carrillo, M. P.;

Dwarakanath, S.; Bhahdigadi, S.; Rao, P.; Kiel, J. L.; Andrews, C. J. J.Fluorescence 2007, 17, 193–199.

(12) Liss, M.; Petersen, B.; Wolf, H.; Prohaska, E. Anal. Chem. 2002, 74, 4488–95.

(13) Li, J. W. J.; Fang, X. H.; Tan, W. H. Biochem. Biophys. Res. Commun. 2002,292, 31–40.

(14) Yamamoto, R.; Baba, T.; Kumar, P. K. Genes Cells 2000, 5, 389–96.(15) Bang, G. S.; Cho, S.; Kim, B. G. Biosens. Bioelectr. 2005, 21, 863–70.

Anal. Chem. 2010, 82, 1851–1857

10.1021/ac9025237 2010 American Chemical Society 1851Analytical Chemistry, Vol. 82, No. 5, March 1, 2010Published on Web 02/01/2010

real-time optical biosensors.8,16-18 The FRET-based aptamerbeacons have been particularly popular.8,11,16,17,19,20 As shown inFigure 1 this sensing scheme involves formation of a duplex wherea fluorophore-labeled aptamer is hybridized with an antisenseoligonucleotide sequence carrying a quencher. The aptamerbeacon shows no fluorescence in duplex; however, the additionof a target analyte results in displacement of the quencher-carryingstrand, disruption of the FRET effect, and the appearance of thefluorescence signal. While a number of aptamer beacons beendescribed in the literature,21,22 to the best of our knowledge, therehave been no reports describing detection of IFN-γ using thissensing strategy.

Given the importance of IFN-γ as a diagnostic immuneresponse marker,1-3,23 we sought to design a novel aptamer-basedimmunosensor for the detection of this analyte. The IFN-γ-bindingDNA aptamer previously described in the literature24,25 wasbiotinylated and immobilized on the surface via avidin-biotininteractions. Surface plasmon resonance (SPR) was used toinvestigate the effects of biotinylation, fluorophore attachment,and spacer incorporation on the ability of aptamer to bind IFN-γ.The 3′ end biotinylated aptamer without spacer was to found tohave the highest affinity for IFN-γ (Kd ) 3.4 nM) and was used

throughout the study. SPR experiments also revealed rapidbinding of IFN-γ molecules with an aptamer/antisense duplexand suggested displacement of an antisense strand by thecytokine molecules. Disruption of the DNA duplex and forma-tion of aptamer-IFN-γ complex was further confirmed withfluorescence assays involving soluble or surface-immobilizedaptamer beacons. To highlight biosensing application of thisapproach, surface immobilization of aptamer beacons anddetection of IFN-γ was demonstrated inside microfluidicdevices.

MATERIALS AND METHODSChemicals and Materials. Glass slides (75 × 25 mm2) were

obtained from VWR (West Chester, PA). 3-Aminopropyl-trimethoxysilane was purchased from Gelest, Inc. (Morrisville,PA). Anhydrous toluene (99.9%), 2-hydroxy-2-methylpropiophe-none (photoinitiator), bovine serum albumin (BSA), HEPES,KCl, EDTA, MgCl2, surfactant Tween20, and glutaraldehydewere obtained from Sigma-Aldrich (St. Louis, MO). Acetonewas obtained from EMD Chemicals (Gibbstown, NJ), Neutra-vidin was purchased from Invitrogen (Carlsbad, CA). Recom-binant human IFN-γ and Interleukin-2 were purchased fromR&D systems (Minneapolis, MN) and Endogen (Woburn, MA),respectively. 96-Well plates, transparent optical flat bottom,black, were purchased from NUNC. 96-Well Reacti-bind neu-troavidin covered plates, black, were obtained from Pierce. Cellculture medium RPMI 1640: 1X, with L-glutamine was pur-chased from VWR.

The following buffers were used in this study: TKM buffer (50mM TrisHCl, 10 mM KCl, 1 mM MgCl2, pH 8.6), HKE buffer(10 mM Hepes, 100 mM KCl, 1 mM EDTA, pH 7.4), HKMTwashing buffer (10 mM Hepes, 100 mM KCl, 1 mM MgCl2, 0.05%Tween20, pH 7.4).

IFN-γ aptamer sequences with 3′ and 5′ biotin modifications(3′B and 5′B) and sequences with biotin + spacer modifications(3′Bspacer and 5′Bspacer) were ordered from Bioneer (Alameda,CA). A 3′ biotinylated aptamer carrying carboxyfluoresceinlabel (FA) and 3′BHQ-1-labeled complementary oligo (Q) weresynthesized by IDT Technologies (San Diego, CA). Oligonucle-otide sequences and modifications used in this study are listedin Table 1.

Prior to their use, samples were heated at 95 °C for 3 min andthen allowed to cool slowly to room temperature. Oligonucleotidesamples were kept overnight at 4 °C until their use. Dilutedsolutions of oligos and recombinant protein for measurementswere prepared in appropriate buffers.

(16) Urata, H.; Nomura, K.; Wada, S.; Akagi, M. Biochem. Biophys. Res. Commun.2007, 360, 459–463.

(17) Nishihira, A.; Ozaki, H.; Wakabayashi, M.; Kuwahara, M.; Sawai, H. NucleicAcids Symp. Ser. (Oxford) 2004, 135–6.

(18) Babendure, J. R.; Adams, S. R.; Tsien, R. Y. J. Am. Chem. Soc. 2003, 125,14716–7.

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

(20) Li, W.; Yang, X. H.; Wang, K. M.; Tan, W. H.; Li, H. M.; Ma, C. B. Talanta2008, 75, 770–774.

(21) Hall, B.; Cater, S.; Ellington, A. D. Biotechnol. Bioeng. 2009, 103, 104901059.(22) Yang, C. J.; Jockusch, S.; Vicens, M.; Turro, N. J.; Tan, W. Proc. Nat. Acad.

Sci. 2005, 102, 17278–17283.(23) Karlsson, A. C., J. N; Martin, S. R.; Younger, B. M.; Bredt, L.; Epling, R.;

Ronquillo, A. V.; Deeks, S. C.; McCune, J. M.; Nixon, D. F.; Sinclair, a. E.J. Immunol. Methods 2003, 283, 141–153.

(24) Lee, P. P.; Ramanathan, M.; Hunt, C. A.; Garovoy, M. R. Transplantation1996, 62 (9), 1297–1301.

(25) Balasubramanian, V.; Nguyen, L. T.; Balasubramanian, S. V.; Ramanathan,M. Mol. Pharmacol. 1998, 53, 926–32.

Figure 1. Schematic representation of an aptamer beacon fordetection of IFN-γ. In duplex, fluorescence of an aptamer is effectivelyquenched by a FRET effect resulting from proximity of fluorophore-labeled aptamer to an acceptor-carrying complementary strand.Binding of IFN-γ disrupts the DNA duplex and results in a fluores-cence signal.

Table 1. Sequences of IFN-γ Aptamer OligonucleotidesUsed in the Experiments

name sequence with modification

5′B 5′-biotin-GGG GTT GGT TGT GTTGGG TGT TGT GT-3′

3′B 5′-GGG GTT GGT TGT GTT GGGTGT TGT GT-Biotin-3′

5′Bspacer 5′-Biotin-C12-GGG GTT GGT TGT GTTGGG TGT TGT GT-3′

3′Bspacer 5′-GGG GTT GGT TGT GTT GGGTGT TGT GT-C12-Biotin-3′

FA 5′-6-FAM- T GGG GTT GGT TGTGTT GGG TGT TGT GT-Biotin-3′

Q 5′- ACAACCAACCCCA-BHQ-1-3′

1852 Analytical Chemistry, Vol. 82, No. 5, March 1, 2010

Characterization of IFN-γ Binding to an ImmobilizedAptamer. SPR experiments were performed on a four-channelBIAcore T3000 instrument (Uppsala, Sweden) using streptavidin(SA) sensor chips obtained from BIAcore. Biotinylated aptamerwas diluted in HKE buffer to 1 µM and injected into SPRinstrument at a flow rate of 20 µL/min. All SPR experiments wereperformed at 20 °C in filtered degassed HKE buffer. One channelof SA sensor chip was designated as reference and was blockedwith biotin (without aptamer) to prevent other binding events fromhappening. IFN-γ solutions ranging in concentration from 12 to120 nM were prepared HKE buffer and injected into the SPRinstrument at flow rate of 20 µL/min. Binding of IFN-γ to theimmobilized aptamer was followed in real-time to determine thetime required for reaching equilibrium. In a typical experiment,the contact time of 180 s was sufficient to reach saturation of thebinding signal and 360 s was allotted for dissociation. Washing inbetween binding steps was performed using HKE buffer. Fouraptamer variants were tested to determine aptamer immobiliza-tion/modification strategy leading to highest binding affinity ofIFN-γ. The variants were: 3′ biotinylated aptamer with or withoutspacer and 5′ biotinylated aptamer with or without spacer. Kd

values were obtained using affinity in solution fitting modelon BIAvaluation 4.0 software.

From the SPR experiments described above, 3′ biotinylatedaptamer without spacer was found to have the lowest Kd and wasused for construction of the DNA duplex-based aptamerbeacon. Hybridization of antisense strand to aptamer and itsdisplacement with IFN-γ molecules were investigated by SPR.In these experiments, aptamer-containing surfaces were primedby injecting 5 µL of 20 mM NaOH at the flow rate of 30 µL/min, followed by washing with HKE running buffer for 1200 s.Hybridization was initiated by injecting 45 µL of 2 µMcomplementary oligonucleotide at 30 µL/min followed by a flowof HKE running buffer at 30 µL/min for 300 s. Equilibriumconstant for aptamer-antisense hybridization was determinedas described in the previous paragraph.

In SPR experiments investigating analyte binding to themolecular beacon duplex, 100 µL of 2 µM quenching oligonucle-otide was injected and flowed at 30 µL/min for 300 s, followed byinjection of 100 µL of 100 nM IFN-γ.

Measuring Fluorescence Signal Due to Aptamer-IFN-γBinding. In addition to characterizing cytokine-aptamer interac-tions with SPR, fluorescence spectroscopy and microscopy wereused to detect changes in fluorescence signal due to analytebinding. The function of surface immobilized aptamer beacon wastested using known concentration dissolved in either buffer orcell culture medium supplemented with 10% serum. Biotinylatedand fluorophore-labeled aptamer was immobilized in neutravidin-coated 96-well plates by incubation of 50 nM aptamer solution for2 h at room temperature. After washing three times with TKMTwashing buffer, quenching-labeled antisense oligo was added ata concentration 500 nM and incubated overnight to allow forhybridization to occur. After washing with TKMT, IFN-γ concen-trations ranging from 1 to 200 nM were prepared either in TKMTbuffer or RPMI medium supplemented with 10% serum and wereadded into the wells of the microplate. Change in fluorescenceintensity was measured with a Safire2 microplate reader (Tecan)at 483 nm excitation and 525 nm emission. The fluorescence signal

was normalized to the background fluorescence of the solutionwithout any input molecules and presented as fold fluorescenceincrease. Several oligos of varying lengths and nucleotide se-quence were tested in terms of quenching efficiency and displace-ment by IFN-γ in order to identify a suitable candidate (see Table1 for the sequence of the complementary strand).

Detecting IFN-γ in Aptamer-Modified Microfluidic De-vices. To demonstrate a proof-of-concept microdevice for detectionof IFN-γ, aptamer beacon molecules were immobilized insideavidin-coated poly(dimethyl siloxane) (PDMS) microchannels.Prior to avidin coating, glass slides were cleaned using “piranha”solution as described by us previously26 and stored in the ovenat 200 °C prior to use. Immediately before silanization, a glassslide was exposed to oxygen plasma for 5 min at 300 W (YES3,Yield Engineering Systems, Livermore, CA) and then placed for10 min in a 2% v/v solution of aminopropyl-triethoxysilane inacetone. After silanization, the slides were rinsed with DI water,dried under nitrogen, cured at 100 °C for 1 h, and incubated in a2% v/v aqueous solution of glutaraldehyde for 1 h.

PDMS microfluidic devices were fabricated using standardSU-8 processing and soft lithography protocols. The design of themicrofluidic devices used in these experiments has been describedin our previous publications.27,28 Briefly, the microfluidic devicecontained two flow chambers with width-length-height dimensionsof 3 × 10 × 0.1 mm and a network of independently addressedauxiliary channels. The auxiliary channels were used to applynegative pressure (vacuum suction) to the PDMS mold and toreversibly secure it on top of a glass substrate. This strategyallowed to seal a fluid conduit on top of the glass slide withoutcompromising the aminosilane layer. The inlet/outlet holes werepunched with a blunt 16 gauge needle. A 5 mL syringe wasconnected to silicone tubing (1/32 in. i.d., Fisher), which wasattached to the outlet of the flow chamber with a metal insert cutfrom a 20 gauge needle. A blunt, shortened 20 gauge needlecarrying a plastic hub was inserted in the inlet. A pressure-drivenflow in the microdevice was created by withdrawing the syringepositioned at the outlet with a precision syringe pump (HarvardApparatus, Boston, MA).

Aminosilane- and glutaraldehyde-modified glass slides wereoutfitted with PDMS microchannels and incubated with 1 mg/mL neutravidin solution in 1× PBS. Biotinylated and fluorescein-labeled aptamer was then injected in the microfluidic channels atconcentration of 10 µM and incubated for 2 h at room temperature.After washing with TKMT buffer, quencher-labeled antisenseoligonucleotide was injected into channels at concentration of 50µM and hybridized with aptamer overnight at room temperature.This step resulted in immobilization of an aptamer-fluorescein/antisense-quencher duplex on the surface of the microfluidicchannels.

During cytokine detection experiments, IFN-γ was injected intothe microfluidic device at concentrations ranging from 1 to 100nM in TKM buffer. The change in fluorescence due to cytokine-aptamer beacon interactions was monitored using a Zeiss 200 M

(26) Jones, C. N.; Lee, J. Y.; Zhu, J.; Stybayeva, G.; Ramanculov, E.; Zern, M. A.;Revzin, A. Anal. Chem. 2008, 80, 6351–7.

(27) Zhu, H.; Macal, M.; Jones, C. N.; George, M. D.; Dandekar, S.; Revzin, A.Anal. Chim. Acta 2008, 608, 186–96.

(28) Zhu, H.; Stybayeva, G.; Macal, M.; Ramanculov, E.; George, M. D.;Dandekarb, S.; Revzin, a. A.; Lab Chip 2008, 8, 2197–2205.

1853Analytical Chemistry, Vol. 82, No. 5, March 1, 2010

epifluorescence microscope (Carl Zeiss MicroImaging, Inc. Thorn-wood, NY) equipped with xioCam MRm (CCD monochrome, 1300pixels × 1030 pixels). Objectives, camera, and fluorescence filterswere computer controlled through a PCI interface. Image acquisi-tion and fluorescence analysis were carried out using AxioVisionsoftware (Carl Zeiss MicroImaging, Inc. Thornwood, NY).

RESULTS AND DISCUSSIONThe goal of this study was to develop an aptamer beacon for

FRET-based detection of IFN-γ. Several key parameters pertainingto orientation of the immobilized aptamer and the design ofaptamer-antisense duplex were characterized by SPR as well asfluorescence spectroscopy and microscopy. As a proof-of-conceptbiosensor demonstration, aptamer beacon molecules were im-mobilized inside microfluidic channels and were shown to producea fluorescence signal in response to different concentrations ofIFN-γ.

Characterization of IFN-γ Binding to Surface ImmobilizedAptamer. Avidin-biotin interactions have been used widely forsurface binding of functional biomolecules, including aptamers;7,29

therefore, this immobilization scheme was chosen for our study.While the nucleic acid sequence of aptamer specific for IFN-γhas been reported in the literature,24 the position of the sensingnucleotides on the DNA strand was not known. Given thatchemical modification may negatively impact the affinity ofaptamer for the analyte,6,30 we investigated the effects of placingbiotin at the 3′ vs 5′ end of the aptamer. In addition, insertion ofPEG spacer between the aptamer and biotin was explored as ameans of making nucleotides more accessible to the target analyte.The aptamer variants including 5′ biotin, 3′ biotin, 5′ biotinw/spacer, and 3′ biotin w/spacer (see Table 1) were synthesizedand immobilized on avidin-coated sensor chips. SPR was used totest the ability of aptamer variants to bind IFN-γ molecules. Arepresentative experiment is shown in Figure 2 where an SPRsensor chip containing four aptamer variants described above waschallenged with 100 nM IFN-γ. As seen from this sensogram, thehighest level of cytokine binding was observed on a 3′ biotinylatedaptamer without spacer. By repeating binding experiments forIFN-γ concentrations ranging from 1 to 100 nM, equilibriumbinding constants Kd for aptamer variants were determinedusing simple affinity fitting model with BIAcore software 4.0.As seen from Table 2, the Kd values ranged from 28 nM in thecase of lowest affinity aptamer biotinylated at the 5′ to 3 nMfor the highest affinity aptamer biotinylated at the 3′. Theseresults show that attachment of biotin at the 5′ end hindersanalyte binding and may mean that the nucleotides responsiblefor recognition of IFN-γ are located at the 5′ end of the aptamer.It is not entirely clear at this time why incorporation of a spacerat the 3′ end increases Kd but may suggest that inclusion ofPEG-based spacers hinders interaction with cytokine moleculesdue to hydration of PEG. Testing other space chemistries willhelp address this question in the future. Overall, the IFN-γaptamer immobilized via 3′ end without the spacer was foundto have the best affinity constant (Kd ) 3 nM) and thereforewas chosen as the basis for the molecular beacon described

in the subsequent sections of this paper. Importantly, Kd valuesof the aptamer-IFN binding were comparable to concentra-tions of cytokine secreted by cells in vitro or observed in

(29) Collett, J. R.; Cho, E. J.; Ellington, A. D. Methods 2005, 37, 4–15.(30) Walter, J. G.; Kokpinar, O.; Friehs, K.; Stahl, F.; Scheper, T. Anal. Chem.

2008, 80, 7372–7378.

Figure 2. SPR analysis of IFN-γ binding as a function of aptamermodification. Vertical arrows represent a washing step where bufferis introduced. (A) Sensograms of four aptamer variants differing inthe placement biotin (B) and inclusion of spacer in addition to biotin(BS). These sensograms compare response four aptamer variantsbinding 100 nM of IFN-γ. Multiple concentrations of IFN-γ were testedfor each aptamer variant to determine Kd values listed in Table 2. (B)SPR sensograms comparing aptamer carrying a fluorophore (F) atthe 5′ end and biotin at the 3′ end to an aptamer without a fluorophoreand biotin 3′ end. These variants were challenged with differentconcentrations of IFN-γ to determine Kd values.

Table 2. Dissociation for Different Aptamer VariantsInvestigated in This Studya

aptamer IFN-γ KD (nM)

5′B 28.1 ± 1.095′B 14.9 ± 1.33′B 3.44 ± 0.23′BS 6.56 ± 0.8F 5.73 ± 1.1

a The abbreviations are as follows: aptamer modified with biotin (B),biotin and spacer (BS), or fluorophore (F).

1854 Analytical Chemistry, Vol. 82, No. 5, March 1, 2010

blood,2,27 pointing to the potential use of aptamer-based sensorfor monitoring physiological levels of IFN-γ.

Design and Characterization of the Aptamer-AntisenseDuplex. The aptamer beacon designs may be broadly categorizedinto monochromophoric and bichromophoric approaches.8 Thefirst strategy is suitable in the case where analyte binding causessignificant structural reorganization of the aptamer leading to achange in the spectroscopic properties of the fluorophore.Fluorescence spectra of 6FAM labeled IFN-γ aptamer were notmuch different before and after IFN-γ binding (data not shown).This result suggests that the binding of IFN-γ did not cause thechange to the aptamer and/or fluorescent properties of thefluorophore. Therefore, we chose to pursue a bichromophoricstrategy involving fluorophore-labeled aptamer and quencher-labeled complementary (antisense) oligonucleotide strand. Asshown diagrammatically in Figure 1, the molecular beacon wascomprised of two DNA oligonucleotides: an aptamer modified witha fluorophore at the 5′ end and an antisense strand labeled withquencher at the 3′ end. The quencher-carrying strand was a 12mersequence complementary to the 5′ region of the aptamer. In theabsence of the target, the two DNA molecules assembled intothe duplex structure where fluorophore and quencher were inclose proximitysallowing for the FRET effect. The displacementof the quencher-carrying oligo strand with the IFN-γ was hypoth-esized to disrupt the quenching of the fluorophore leading to afluorescence signal.

In order for disruption of the DNA duplex by the target analyteto occur rapidly, the affinities of aptamer-antisense andaptamer-IFN-γ needed to be similar. SPR experiments were firstcarried out to determine the Kd value for an aptamer modifiedwith fluorescein at the 5′ and biotin at the 3′ end. The affinityconstant for IFN-γ binding to this aptamer construct wasdetermined to be 5.73 ± 1.1 nM, suggesting that fluorophoreattachment did not appreciably impact the binding IFN-γ (seeTable 2 for comparison of Kd values for different aptamermodifications). SPR was also used to determine the equilibriumbinding constant for the hybridization of the quencher-antisensestrand and a fluorophore-aptamer construct (SPR sensogramsnot shown). The Kd value for this interaction was determinedto be 1.09 ± 0.4 nM. The similarity of Kd values foraptamer-IFN-γ and aptamer-antisense interactions suggestedthat displacing the antisense strand in DNA duplex by thecytokine molecules was indeed possible.

Specificity is one of the most important characteristics of abiosensor. SPR experiments were used to demonstrate that ouraptamer was responding specifically to IFN-γ. Figure 3 shows arepresentative experiment where two SPR channels containingaptamer are challenged with 100 nM concentration of IFN-γ andIL-2. As seen from the data, the binding signal was observed onlyin the channel containing the correct analytessuggesting specific-ity of the aptamer. Further proof of aptamer beacon specificity inserum is presented in the next section.

In order to verify competitive binding of IFN-γ, we performedadditional SPR and fluorescence spectroscopy experiments. Arepresentative SPR sensogram is shown in Figure 4. In thisexperiment, one channel was coated with aptamer while the otherchannel contained aptamer-antisense duplex. Importantly, noappreciable dissociation of the DNA duplex was observed after

40 min in a running buffer. Injection of 100 nM IFN-γ resulted inrapid appearance of binding signals of comparable magnitude inboth channels (see Figure 4). The response time (time to 90% ofsignal) for IFN-γ binding was 15 min for both channels. Thesedata were suggestive of displacement of an antisense strand andbinding of IFN-γ to the aptamer; however, there remained apossibility that IFN-γ attached to the DNA duplex withoutdislodging the quencher oligo strand. Fluorescence spectroscopyand microscopy experiments described in the following sectionwere conducted to exclude this scenario.

Quantifying Fluorescence Response of Surface Immobi-lized Aptamer Beacon. To conclusively demonstrate displace-ment of the quenching oligo strand by the cytokine molecules,the biotinylated aptamer-antisense duplex was immobilized inavidin-coated 96-well plates. IFN-γ was then added into the platesat concentrations of 1, 5, 10, 50, and 100 nM and the fluorescenceintensity was measured after 10 min of incubation. This time waschosen based on the SPR studies of the dynamics of IFN-γ to theaptamer-antisense complex described in the previous section.The aptamer beacon response was quantified using a microplatereader. As shown in Figure 5, the fluorescence signal change ofour aptamer beacon was linear from 5 to 100 nM IFN-γ and thelowest detected concentration of analyte was 5 nM. The detectionlimit of our aptamer sensor is sufficient to monitor physiologicallevels of the cytokine secreted by the immune cells.2,27

Our laboratory is interested in placing immunosensors at thesite of the cells in order to characterize dynamics of cytokinerelease;28,31 therefore, we sought to characterize response of theaptamer beacon in cell culture media. In these experiments, IFN-γwas dissolved in RPMI (media commonly used for culturingimmune cells) supplemented with 10% fetal bovine serum or inHEPES buffer (pH 7.4) were compared. As shown in Figure 5,aptamer beacons remained functional in the cell culture mediaand showed concentration dependent changes in fluorescencesignal. This result is very important as it demonstrates that theaptamer beacon remains responsive in a sample where concentra-tion of extraneous proteins exceeds IFN-γ concentration by

(31) Zhu, H.; Stybayeva, G. S.; Silangcruz, J.; Yan, J.; Ramanculov, E.; Dandekar,S.; George, M. D.; Revzin, A. Anal. Chem. 2009, 81, 8150–8156.

Figure 3. Specificity of aptamer-IFN-γ interaction: SPR sensogramshowing binding curves of aptamer-modified surfaces challenged with100 nM IFN-γ vs 100 nM IL-2. No signal is observed for IL-2 binding.Arrows represent a washing step.

1855Analytical Chemistry, Vol. 82, No. 5, March 1, 2010

100-10 000 fold. An ∼20 to 30% loss of signal for aptamer beaconsoperating in serum-containing media may be attributed to low levelnonspecific binding of serum components. Despite this, our resultsare highly encouraging as they demonstrate reproducible androbust responses of an aptamer beacon in a physiological mediumand point to immediate applicability of this sensing strategy forreal sample analysis.

Biosensing applications frequently require integration of therecognition molecules into microdevices to enable analysis of smallsample volumes. In this paper, we demonstrate modification ofthe microfluidic device with aptamer beacon molecules and in situdetection of IFN-γ binding. Because PDMS is solvated easily byorganic solvent, we chose to first modify the glass slides withaminosilane and glutaraldehydesmaking these substrates protein

reactivesand then to place PDMS channels on top. The PDMSfluidic conduits were effectively sealed on glass by using negativepressuresan approach first described by Schiff et al. and em-ployed by us in previous studies.28,31,32 The fluidic channels withprotein-reactive glass surfaces were treated with neutravidin andthen incubated with fluorophore-labeled and biotinylated aptamermolecules. Figure 6A shows a microfluidic device with twochannels where the lower channel contains aptamer-fluorophorewhile the upper channel has been quenched by introduction ofthe quencher-carrying antisense strand. This image, as well ascorresponding fluorescence intensity measurements seen inFigure 6B, demonstrates that injection of a quenching strand intothe fluidic channel decreased the fluorescence by ∼10-fold. Thisis suggestive of DNA duplex formation and effective FRETquenching in the microdevice. Importantly, injection of 10 nMIFN-γ into the channel and subsequent displacement of thequenching strand caused reappearance of the fluorescence signal(Figure 6C). The signal observed in the microfluidic channel wasalso a function of analyte concentration so that introduction of100 nM resulted in higher fluorescence compared to 10 nM ofIFN-γ (Figure 6D; see also Figure 6E). The results shown inFigure 6 are highly significant as they demonstrate integrationof aptamer beacon molecules into a microdevices and in situdetection of IFN-γ.

Our study describes an aptamer beacon for detection of IFN-γsan important clinical indicator of immune function. In contrastto traditional approaches employing antibodies and sandwichimmunoassays which require multiple washing steps and serveas an end-point measurement, the biosensor described here emitsfluorescence signal directly upon binding of the cytokine mol-ecules. This surface immobilized aptamer beacon provides asimple, one-step immunoassay and may therefore be used fordynamic monitoring of cytokine release. The detection limit ofour aptamer beacon (nanomolar range) is not as low as in thework by Min et al. who used impedance spectroscopy to detect

(32) Schaff, U. Y.; Xing, M. M.; Lin, K. K.; Pan, N.; Jeon, N. L.; Simon, S. I. LabChip 2007, 7, 448–56.

Figure 4. SPR sensogram demonstrating IFN-γ binding to an aptamer beacon. Aptamer molecules were immobilized in two channels of anSPR instrument. A quenching strand was injected into channel 1 forming an aptamer-quencher duplex. Channel 2 contained only aptamer andwas used as a reference. In the next step, we injected 100 nM of IFN-γ into both channels and observed comparable binding signals in bothchannels. This suggested disruption of a DNA duplex and displacement of the quenching strand with IFN-γ.

Figure 5. Fluorescence signal from an aptamer beacon challengedwith varying concentrations of IFN-γ. Analyte was dissolved in eitherHEPES buffer (pH 7.4) or RPMI1640 media supplemented with 10%serum. Aptamer beacon responses to varying concentrations of IFN-γwere measured using a microplate reader. The measured fluores-cence intensities were normalized by the values obtained in theabsence of analyte molecules. Data are averages of three indepen-dent experiments (n ) 3).

1856 Analytical Chemistry, Vol. 82, No. 5, March 1, 2010

pM level IFN-γ binding to aptamer interactions.33 This discrepancyis likely due to the fact that ac impedance measures nonequilib-rium molecular binding events at the concentrations far belowthe Kd. For example, the van Bennekom group working withantibody-containing surfaces reported detection of attamolarlevels of IFN-γ using impedance34 and nanomolar levels withSPR.35

CONCLUSIONSThis paper describes development of an aptamer beacon for

FRET-based detection of IFN-γ. SPR was used to establishequilibrium binding constants (Kd) for different aptamer variantsin order to determine how aptamer modification with biotinand fluorophore molecules affected analyte binding. Thesestudies revealed that biotinylation of aptamer at the 3′ endresulted in the lowest Kd of 3.44 nM. Fluorescence spectroscopyand microscopy were used to demonstrate that attachment ofIFN-γ to an aptamer beacon duplex resulted in fluorescence

signal that changed in analyte concentration dependent fashion.The appearance of fluorescence signal suggested displacementof the quenching strand and disruption of the FRET effect, thusvalidating function of the aptamer sensor. To highlight pos-sibility of sensor miniaturization, aptamer beacon moleculeswere immobilized inside microfluidic channels and were shownto be responsive in situ to different IFN-γ concentrations. Thepossibility for direct and dynamic sensing of cytokine bindingprovided by this aptamer beacon will be leveraged in the futurefor detecting cell-secreted cytokines in real-time.

ACKNOWLEDGMENTWe thank Prof. Laura Marcu and Dr. Yinghua Sun in the

Department of Biomedical Engineering at UC Davis for use offluorescence microscope. NT was supported by a fellowship fromthe National Center for Biotechnology, Republic of Kazakhstan.These studies were supported by an NSF EFRI grant awarded toAR.

Received for review November 4, 2009. Accepted January20, 2010.

AC9025237

(33) Min, K.; Cho, M.; Han, S.-Y.; Shim, Y.-B.; Ku, J.; Ban, C. Biosens. Bioelectr.2008, 23, 1819–1824.

(34) Dijksma, M.; Kamp, B.; Hoogvliet, J. C.; Van Bennekom, W. P. Anal. Chem.2001, 73, 901–907.

(35) Stigter, E. C. A.; de Jong, G. J.; van Bennekom, W. P. Biosens. Bioelectr.2005, 21, 474–482.

Figure 6. Immobilization of aptamer beacons in microfluidic devices. (A) Image showing two fluidic channels where the bottom channel containsonly aptamer-fluorophore while the upper channel contains aptamer-fluorophore-quencher duplex. (B) Fluorescence microscopy characterizationof quenching observed in part A. An ∼10 fold quenching was observed. (C-E) Response of the fluidic channels to low (C) and high (D)concentrations of IFN-γ, and the corresponding fluorescence intensity measurement (E).

1857Analytical Chemistry, Vol. 82, No. 5, March 1, 2010