422 © 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Novel siRNA-based molecular beacons for dual imaging and therapy

rely on a fluorescence resonant energytransfer (FRET) fluorophore pair togenerate a signal upon binding tocomplementary sequences [8]. Stud-ies have shown that MBs can discrim-inate between targets that differ byonly a single nucleotide [9]. Used in awide range of biomedical applica-tions, MBs offer the capacity to moni-tor gene expression in single livingcells that more traditional forms ofanalysis do not [10].

We report a proof-of-principledemonstration of siRNAs modifiedinto an siRNA-based MB. A potentialprobe for cancer detection and thera-py was synthesized by creating ansiRNA-based MB that detects and tar-gets the hTR sequence for telomerase[11]. Expressed in over 85% of all tu-mors [12], telomerase is essential fortumorigenesis by preventing celldeath through extending existingtelomeres using an RNA template lo-cated within the enzyme [13–15].siRNAs have been demonstrated aspotential therapeutic agents to treatcancer by silencing the expression oftelomerase and inducing cancer cellsenescence and apoptosis [11, 16].

IntroductionSmall interfering RNAs (siRNAs) arecomposed of complementary RNAstrands of 21–23 nucleotides (nts) inlength with sense and antisenseorientation targeting the mRNA of in-terest [1, 2]. Shown as powerfulmethods of gene silencing with highspecificity, RNA interference (RNAi)is mediated by these siRNAs, whichdirect mRNA target cleavage basedon Watson-Crick base pairing of theguide strand (antisense) in the RNA-induced silencing complex (RISC)[3–5].

The discovery of RNAi has gener-ated tremendous interest by bothbiotechnology and pharmaceuticalcompanies to evaluate siRNA applica-tions ranging from animal knockouts[6] to treatment of human diseases [7],such as age-related macular degener-ation [6] or HIV by targeting viral repli-cation mechanisms [7]. Presently,siRNA has solely been used therapeu-tically, while siRNA imaging has beenlimited to end-labeling with fluo-rophores to monitor delivery of siRNAinto cells. Our hope is to leverageRNAi towards diagnostic and moni-toring applications of mRNA by utiliz-ing molecular beacons (MBs).

MBs are highly sensitive and spe-cific fluorescent nucleic acid probesused for monitoring the expression ofnucleic acids at the single-cell level.Consisting of four key components – aloop, stem (typically 4–6 bases), adonor and acceptor fluorophore, MBs

Short interfering RNAs (siRNAs) have become a mainstream tool reliably used to studyand silence protein expression. We offer a proof-of-principle demonstration thatsiRNAs may be modified into a siRNA-based molecular beacon that activates uponbinding to sequence-specific mRNA in cells while mediating RNA interference. We suc-cessfully demonstrate detection and knockdown of telomerase expression in humanbreast cancer cells. This probe provides a novel look at siRNA target validation that isnot currently possible in live cells and holds promising potential in biological applica-tions for disease detection and therapy based on mRNA expression, such as a telom-erase-targeted siRNA probe in cancer.

We synthesized our novel siRNA-based probe utilizing two short 21-ntRNA complementary strands as thestem and a highly flexible poly(ethyl-ene glycol) (PEG) molecule covalentlylinked to the RNA strands as the loopof an siRNA-based MB (Fig. 1). Twodifferent PEG lengths were used,Mr 3400 and 5000. A commonly usedFRET fluorophore pair, Cy3 and Cy5fluorophores, was conjugated to the 3´terminus of the antisense and 5´ ter-minus of the sense strand, respective-ly. Cy3 and Cy5 emission peaks are at565 and 667 nm. Previous studieshave shown that modifications of alltermini of the siRNA duplex exceptthe 5´ terminus of the antisense strandcan be made without significantly af-fecting gene silencing [17]. After thesiRNA probe incorporates into RISC,the anti-guide strand would presum-ably be cleaved at the ninth position[5], leaving a nine-base pairing stemmolecular probe. Longer MB stemshave shown to have improved abilityto discriminate targets and yield high-er signal-to-background ratios thanshorter stems [18]. Next the presenceof a complementary mRNA would dis-place the anti-guide strand, resultingin activation of the probe for imaging.

siRNA-based probe synthesisModified 21-mer oligoribonucleotide(ss/siRNAs) were purchased commer-cially (Dharmacon RNA Technologies,Lafayette, CO, USA) utilizing a previ-ously published and validated se-quence targeting the telomere repeat

BTJ

-FO

RU

MBiotechnology

Journal DOI 10.1002/biot.200600257 Biotechnol. J. 2007, 2, 422–425

Forum Contribution

Figure 1. siRNA-basedMB in comparison tothe conventional MB.

422_200600257_ForumContrib.qxd 02.04.2007 10:01 Uhr Seite 422

© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim 423

template sequence (hTR) [11]. Themodified sense strand sequence is: 5´HS-U*UGUCUAACCCUAACUGAG-TT-3´ where U* is a Cy5-dUTP fluo-rophore-conjugated base. The modi-fied antisense strand is: 3´ NH2-T*T-AACAGAUUGGGAUUGACUC-5´where T* is a Cy3-dTTP fluorophore-conjugated base. Cy3 and Cy5 are acommonly used fluorophore FRETpair. The sense and antisense oligori-bonucleotides were reacted with amaleimide-PEG-NHS ester heterobi-functional linker (3400 and 5000 MrPEG; Nektar Therapeutics, Huntsville,AL, USA). During the coupling reac-tion, siRNA antisense and sensestrands were added in a twofold molarexcess to ensure sufficient couplingonto the heterobifunctional PEG link-er. The reaction was allowed to pro-ceed overnight in the dark in RNase-free phosphate-buffered saline, pH 7.4at 4°C.

siRNA-based probe isolationThe conjugated ssRNA-PEG-ssRNAwas then isolated from free ssRNAand PEG by 15% denaturing PAGE.The sample was heated at 90°C for2 min prior to gel loading, mixed withgel loading buffer, and electro-phoresed at 35 mA. The PAGE gel waspost-stained with a 0.5 μg/mL ethidi-um bromide solution. Two distinctbands were seen for each probe. Areference nucleotide sizing marker(Decade Markers, Ambion, Austin,TX) was run in parallel with the bandsto determine which band containedthe conjugated siRNA probe. ThesiRNA probe band (determined by nu-cleotide size) was then carefully ex-cised from the gel and isolated by thecrush and soak method. After 48-helution at 4°C in the dark with 10 mL1 M NaCl, the sample was centrifugedat 2000 × g for 5 min. The elutionbuffer was ethanol precipitated with a4-volume excess of cold 90% ethanol/10% 3 M sodium acetate at –20°C for48 h to isolate the siRNA probe. Thesolution was centrifuged at 20 000 × gfor 30 min at 4°C and resuspended inannealing buffer (20 mM KCl, 6 mMHEPES-KOH pH 7.5). Annealing was

performed by heating to 60°C for5 min, and slow-cooled to roomtemperature over 30 min. Extinct-ion measurements at 260 nm (ε =414 600 L/mol cm) were performed ina quartz cuvette to determine finalsiRNA conjugate concentration on aVarian Cary 50 Bio spectrophotometer(Walnut Creek, CA, USA). The probewas then aliquoted into RNase-freetubes and stored at –80°C.

Spectroscopy measurementsAll measurements were performedwith a constant temperature 1.5-mLstoppered quartz fluorescence cu-vette (Starna Cells, Atascadero, CA,USA) on a Horiba Jobin Yvon SPEXFL3-22 Fluorimeter (Edison, NJ, USA)with dual excitation and emissionmonochromators. Sample tempera-ture was controlled by a circulatingtemperature water bath through thequartz cuvette. Time-integrated pho-toluminescence was measured using488 nm excitation light and emissionscans from 525 to 800 nm. Bandpassslits and integration time were set to3 nm/3 nm and 100 ms, respectively,on the fluorimeter. All values were nor-malized over time to a rhodamine 6Gstandard to avoid any artifacts thatcould arise from possible lamp fluctu-ations. FRET efficiency between Cy3and Cy5 fluorophores was calculatedby intensity of Cy5 fluorescence/(in-tensity of Cy3 fluorescence + intensi-ty of Cy5 fluorescence).

Cell culture studiesThe human breast cancer cell line SK-BR-3 and normal breast fibroblast cellline CCD-1059Sk were from AmericanType Culture Collection (ATCC, Man-assas, VA, USA). SK-BR-3 cells werecultured (37°C, 5% CO2) in McCoy’s5A medium with 10% fetal bovineserum (Invitrogen Corp., Carlsbad,CA, USA). CCD-1059Sk cells were cul-tured (37°C, 5% CO2) in minimum es-sential medium (Eagle) with 10% fetalbovine serum. Both SK-BR-3 cells andCCD-1059Sk cells were plated ontoglass chamber slides (Nalge Nunc In-ternational, Rochester, NY, USA) formicroscopy studies. Cells were plated

into six-well tissue culture plates at aconcentration to provide 80% conflu-ence 24 h later prior to transfection us-ing Lipofectamine 2000 (Invitrogen).At that time, the siRNA probe was di-luted to a final volume of 500 μL withOpti-Mem Reduced Serum Media. Ina separate tube, 30 μL Lipofectaminewas diluted with 470 μL Opti-Mem.The two solutions were separatelymixed gently and incubated at roomtemperature for 5 min. The contentswere then combined and mixed gen-tly by pipetting and incubated at roomtemperature for 30 min. The liposomecomplexes were then added to theculture medium and mixed gently for30 s. After 44 hours, the cells weretrypsinized, counted, and 2000 cellsremoved for assay of telomerase activ-ity. Cell studies were performed withthe 3400-Mr siRNA probe, 5000-MrsiRNA probe, non-targeting SilencerNegative Control no. 1 siRNA (Am-bion, Austin, TX), and mock transfec-tion at 100 nM siRNA concentrations.

siRNA imagingAll images were collected using aZeiss LSM 510 META NLO confocalsystem mounted on an Axiovert 200Minverted microscope and Plan-Apo-chromat 63× objective lens (Carl ZeissMicroimaging, Inc., Thornwood, NY,USA). siRNA probes were excitedwith a 488-nm Argon laser source inlive cell imaging. Spectrofluorimetermeasurements demonstrated mini-mal excitation of the Cy5 fluorophorewith a 488-nm excitation source, indi-cating that Cy5 emission generationis due to fluorescence resonance ener-gy transfer from the Cy3 fluorophore.All images were acquired with thesame detector, gain, pinhole, andpower settings at 1024 × 1024 pixels toallow direct visual comparison be-tween normal and cancer cells. Lamb-da emission scanning from 510 to700 nm was performed using theMETA detector to verify fluorescenceintensity changes between Cy3 andCy5 fluorophores.

Biotechnol. J. 2007, 2, 422–425 www.biotechnology-journal.comB

TJ-FOR

UM

422_200600257_ForumContrib.qxd 02.04.2007 10:02 Uhr Seite 423

424 © 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Telomerase activityTelomerase activity was assayed us-ing a commercial fluorescence-basedTRAPEZE® XL telomerase detectionkit (Intergen, Purchase, NY, USA),which utilizes a modified telomeric re-peat amplification protocol. Lysates(1000 cell-equivalents) were mixedwith the TRAPEZE XL reaction mixcontaining Amplifuor™ primers, andincubated at 30°C for 30 min. Samplesunderwent PCR amplification. Ampli-fied telomerase products were quanti-tated with a SpectraMax M2 fluores-cent plate reader. Telomerase activitywas calculated by comparing the ratioof telomerase products to an internalstandard for each lysate, as describedby the manufacturer.

Results and discussionAfter synthesis and isolation of theprobe, we evaluated the hybridizationof the siRNA-based probes upon addi-tion of complementary target strandsin vitro. After a 20-min incubation at30°C, a 9.3-fold and 8.2-fold decreasein Cy5 fluorescence occurred for 3400-and 5000-Mr siRNA probes, respec-tively. In addition, temperature cy-cling studies between 15 and 90°C

were performed to examine the sepa-ration and re-annealing of the siRNAprobe based on FRET efficiency (seeSupporting information).

We next evaluated siRNA probetargeting to telomerase in live humancells. Cancer and normal cells weretransfected with 100 nM siRNA probeusing cationic lipid-based transfec-tion agents. The sequence of our si-RNA probe targeted the hTR se-quence of telomerase mRNA. Cancercells express high levels of telomerasemRNA, while normal cells minimallyexpress telomerase mRNA. Confocalimaging was performed every 2 hafter transfection up to 14 h and thenevery 4 h up to 38 h post-transfection.Figure 2 demonstrates the activationof the siRNA probe in cancer cells incontrast to normal cells. Interestingly,the 3400-Mr siRNA probe discriminat-ed differences between normal andcancer cells earlier than the 5000-MrsiRNA probe. The ideal time for imag-ing the siRNA probes appeared tobe 6–10 h post-transfection for the3400-Mr siRNA probe and 10–14 hpost-transfection for the 5000-MrsiR-A probe. At 24 h post transfection,no difference was observed between

normal and cancer cells (no FRET ob-served), presumably due to completeloss of the siRNA probe stem bydegradation from intracellular exonu-cleases.

Lastly, we evaluated the efficacy ofthe siRNA-based probe at mediatingRNAi. A commercial telomerase activ-ity assay was performed on the cells44 h post transfection with the siRNAprobe. Cells appeared healthy aftertransfection prior to telomerase assay.Figure 3 demonstrates effective genesilencing of telomerase by the siRNA-based probe. This finding is consis-tent with previous observations oftelomerase silencing using this siRNAprobe sequence [11]. More important-ly, the modifications to siRNA did notinhibit effects of gene silencing inagreement with previous studies [17].

By tethering the anti-guide strandof the siRNA duplex to the guidestrand, the two strands are able to re-anneal if no mRNA is present to com-pete for hybridization. Whereas thevast majority of MBs are constructedusing an oligonucleotide backbonecomprising the loop and stem, it wasnecessary to utilize a synthetic PEGlinker as the loop for our design sinceDicer, an RNAi protein, has beenshown to cleave oligonucleotide hair-pin beacons such as shRNAs into twoseparate strands during RNAi [17].Future work includes evaluating dif-ferent siRNA sequences to further ex-amine specificity and off-targeting ef-fects of these modified siRNA probesas well as imaging and silencing ofother proteins and improving existingfluorophore selection for in vivo imag-ing modalities. Potential future appli-cations of this method could be its usein studying diseases or monitor RNAi.Ultimately, rapid delivery of siRNAprobes into cells could provide a use-ful clinical tool in early detection andtherapy of diseases.

Although there have been a sub-stantial number of recent studies eval-uating the therapeutic applications ofsiRNA, we describe here a novel bio-engineering modification to siRNAs.In this study, we provide an initialproof-of-principle demonstration that

BTJ

-FO

RU

MBiotechnology

Journal Biotechnol. J. 2007, 2, 422–425

Figure 2. Confocal images at 10 and 14 h post transfection with Lipofectamine 2000 in normal humanbreast fibroblast and human breast cancer cells transfected with 3400- and 5000-Mr siRNA-basedprobes targeted against the hTR sequence in telomerase. Microscopy settings were held constant forimage comparison. Analysis of the images shows a significant increase in Cy5 fluorescence due toFRET in normal cells that do not express telomerase versus cancer cells that do express telomerase.

422_200600257_ForumContrib.qxd 02.04.2007 10:02 Uhr Seite 424

© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim 425

siRNAs may be modified into a fluo-rescence-based MB for mRNA detec-tion and silencing. Using siRNA-based probes in the imaging and si-lencing of telomerase could potential-ly be useful in detecting and treatingprecancerous lesions. These modifiedsiRNAs may provide promising newways to evaluate mRNA expression indiseases and serve as useful re-porter/gene silencing tools for targetvalidation in basic science and fordual imaging and therapy in clinicalresearch.

This work was supported by theNanoscale Science and EngineeringInitiative of the National ScienceFoundation under NSF AwardNumber EEC-0118007. We thankStephanie Greene for her assistancewith gel electrophoresis, Christie Pee-bles for her assistance with PCR, andthe Dharmacon Production Staff, par-ticularly Rob Kaiser, for oligonu-cleotide synthesis.

by Emmanuel Chang, Ming-Qiang Zhuand Rebekah DrezekRice University, Department of Bio-engineering, Houston, Texas, USAE-mail: Drezek@rice.edu

References[1] Hamilton, A., Baulcombe, D., A species of

small antisense RNA in posttranscriptionalgene silencing in plants. Science 1999, 286,950–952.

[2] Elbashir, S. M., Harborth, J., Lendeckel, W.,Yalcin, A. et al., Duplexes of 21-nucleotideRNAs mediate RNA interference in cul-tured mammalian cells. Nature 2001, 411,494–498.

[3] Hannon, G. J., RNA interference. Nature2002, 418, 244–251.

[4] Liu, J., Carmell, M. A., Rivas, F. V., Marsden,C. G. et al., Argonaute2 is the catalytic en-gine of mammalian RNAi. Science 2004,305, 1437–1441.

[5] Rand, T. A., Petersen, S., Du, F., Wang. X.,Argonaute2 cleaves the anti-guide strandof siRNA during RISC activation. Cell 2005,123, 621–629.

[6] McCaffrey, A. P., Meuse, L., Pham, T. T.,Conklin, D. S. et al., Gene expression: RNAinterference in adult mice. Nature 2002,418, 38–39.

[7] Shankar, P., Manjunath, N., Lieberman, J.,The prospect of silencing disease usingRNA interference. JAMA 2005, 293,1367–1373.

[8] Tyagi, S., Kramer, F., Molecular beacons:probes that fluoresce upon hybridization.Nat. Biotechnol. 1996, 14, 303–308.

[9] Bonnet, G. S., Tyagi, S., Libchaber, A.,Kramer, F. R., Thermodynamic basis of theenhanced specificity of structured DNAprobes. Proc. Natl. Acad. Sci. USA 1999, 96,6171–6176.

[10] Sokol, D., Zhang, X., Lu, P., Gewirtz, A., Realtime detection of DNA RNA hybridizationin living cells. Proc. Natl. Acad. Sci. USA1998, 95, 11538–11543.

[11] Kosciolek, B., Kalantidis, K., Tabler, M.,Rowley, P., Inhibition of telomerase activityin human cancer cells by RNA interference.Mol. Cancer Ther. 2003, 2, 209–216.

[12] Shay, J., Bacchetti, S., A survey of telom-erase activity in human cancer. Eur. J. Can-cer 1997, 33, 787–791.

[13] Greider, C., Blackburn, E., A telomeric se-quence in the RNA of Tetrahymena telom-erase required for telomere repeats synthe-sis. Nature 1989, 337, 331–337.

[14] Morin, G., The human telomere terminaltransferase enzyme is a ribonucleoproteinthat synthesizes TTAGGG repeats. Cell1989, 59, 521–529.

[15] Feng, J., Funk, W., Wang, S., Weinrich, S. etal., The RNA component of human telom-erase. Science 1995, 269, 1236–1241.

[16] Shammas, M. A., Koley, H., Batchu, R. B.,Bertheau, R. C. et al., Telomerase inhibitionby siRNA causes senescence and apopto-sis in Barrett’s adenocarcinoma cells:mechanism and therapeutic potential. Mol.Cancer 2005, 4, 24–38.

[17] Dorsett, Y., Tuschl, T., siRNAs: Applicationsin functional genomics and potential astherapeutics. Nat. Rev. 2004, 3, 318–329.

[18] Tsourkas, A., Behlke, M. A., Rose, S. D.,Bao, G., Hybridization kinetics and thermo-dynamics of molecular beacons. NucleicAcids Res. 2003, 31, 1319–1330.

Biotechnol. J. 2007, 2, 422–425 www.biotechnology-journal.comB

TJ-FOR

UM

Figure 3. Telomeraseactivity relative to un-treated SK-BR-3 humanbreast cancer cells aftertransfection with100 nM siRNA probeat 44 h (n=3). siRNAprobe is able to mediateRNA interference andsignificantly knockdownexpression of telo-merase in cancer cells.

This contribution was peer reviewedReceived 3 July 2006Revised 27 November 2006Accepted 22 January 2007

422_200600257_ForumContrib.qxd 02.04.2007 10:02 Uhr Seite 425