Analytica Chimica Acta 813 (2014) 35–40
Contents lists available at ScienceDirect
Analytica Chimica Acta journal homepage: www.elsevier.com/locate/aca
Magnetic bead-based hybridization assay for electrochemical detection of microRNA Martin Bartosik a,b,∗ , Roman Hrstka a , Emil Palecek a,b , Borivoj Vojtesek a a b
Regional Centre for Applied Molecular Oncology (RECAMO), Masaryk Memorial Cancer Institute, Zluty kopec 7, 656 53 Brno, Czech Republic Institute of Biophysics, Academy of Sciences of the Czech Republic, v.v.i., Kralovopolska 135, 612 65 Brno, Czech Republic
h i g h l i g h t s
g r a p h i c a l
a b s t r a c t
• Simple and quick electrochemical assay for microRNA detection was developed. • Specific electroactive complex based on osmium(VI) was used for microRNA 3 -end labeling. • Good selectivity was obtained by discriminating perfect duplex and single mismatches. • Labeled microRNA could be detected in total RNA samples.
a r t i c l e
i n f o
Article history: Received 12 November 2013 Received in revised form 6 January 2014 Accepted 8 January 2014 Available online 16 January 2014 Keywords: MicroRNA Electrochemistry Mercury electrodes Mismatch discrimination
a b s t r a c t Aberrant expression of microRNAs (miRNAs), short non-coding RNA molecules regulating gene expression, is often found in tumor cells, making the miRNAs suitable candidates as cancer biomarkers. Electrochemistry is an interesting alternative to current standard methods of miRNA detection by offering cheaper instrumentation and faster assays times. In this paper, we labeled miRNA in a quick, simple, two-step procedure with electroactive complex of osmium(VI) and 2,2 -bipyridine, Os(VI)bipy, which specifically binds to the ribose at the 3 -end of the miRNA, and hybridized such labeled miRNA with biotinylated capture probe attached to the streptavidin magnetic beads. Labeled miRNA was then detected at hanging mercury drop electrode at femtomole level due to an electrocatalytic nature of the peak from the Os(VI)bipy label. We obtained good selectivity of the assay using elevated hybridization temperatures for better discrimination of perfect duplex from single and double mismatches. After optimization of the protocol, we demonstrated feasibility of our assay by detecting target miRNA in real total RNA samples isolated from human cancer cells. © 2014 Elsevier B.V. All rights reserved.
1. Introduction A boom in microRNA (miRNA) research can be observed in past few years, especially due to the finding that altered expression levels of miRNAs in cells are associated with onset and progress of the cancer, making them attractive not only as diagnostic biomarkers, but also as potential therapeutic targets [1]. The link between miRNAs and many diseases is due to their binding to mRNA and
∗ Corresponding author at: RECAMO. Tel.: +420 543 133 325. E-mail address: [email protected] (M. Bartosik). 0003-2670/$ – see front matter © 2014 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.aca.2014.01.023
subsequent negative regulation of gene expression, including many oncogenes and tumor suppressor genes. Most-widely employed method for detection of aberrantly expressed miRNAs is qRT-PCR, where the reverse transcription (RT) first converts target miRNA into the cDNA, followed by its amplification using specifically designed primers and by quantification of the amplified duplexes [2]. The method has a good accuracy and sensitivity, enabling determination of low levels of miRNAs, but the amplification step is not straightforward because miRNAs are usually much shorter than primers. Later, microarrays were introduced for high-throughput and rapid multiplexed detection of miRNAs, both with and without PCR amplification. They are based
36
M. Bartosik et al. / Analytica Chimica Acta 813 (2014) 35–40
on nucleic acid hybridization between fluorescently labeled miRNAs and solid substrate-attached capture probes, followed by a fluorescence read-out of the label. Recently, profiling of miRNAs by next-generation sequencing (NGS) has progressed rapidly. Several NGS platforms were established allowing not only measurement of expression levels of known miRNAs, but also detection of unknown miRNAs [3]. These robust techniques, however, bear higher costs due to sophisticated instruments and demand for highly skilled personnel. Therefore, new methods for miRNA detection are being developed which would be fast, inexpensive and relatively simple [4,5]. Such requirements are met by electrochemical (EC) techniques, which besides above-mentioned advantages also offer possibility for miniaturization and parallel measurements of multiple samples at electrode chips. More and more papers appear each year also for EC-based detection of miRNAs, both in label-free [6–8] and label-based format [9–14]. The latter one, although more complex, usually provides higher sensitivity and versatility by introducing e.g. specific antibodies (conjugated with enzymes and detecting enzymatic products) [9–12], or nanoparticle tags [13,14]. Recently, we have used an electroactive label based on sixvalent osmium, Os(VI), in complex with nitrogenous ligand (L), Os(VI)L, for specific modification of 3 -terminal ribose in RNA [15,16]. This complex is similar to the eight-valent Os(VIII)L, which was developed already in 1980’s for probing the DNA structure by binding to unpaired pyrimidine residues in single-stranded portions of DNA [17]. Labeling of miRNAs with these Os(VIII) complexes would be possible but modification of bases (uracil and cytosine) would interfere with the miRNA hybridization necessary for determination of specific miRNAs. This is not the case with Os(VI)L complexes which were shown to react specifically with diol groups in sugar residues in oligo- and polysaccharides [18–21] as well as in ribosides [22], but not in deoxyribosides (lacking the diol group at the 3 -end of the molecule) or with nucleic acid bases. We mostly relied on 2,2 -bipyridine (bipy) as a ligand for the Os(VI) because Os(VI)bipy-modified adducts yield redox couples at carbon and highly sensitive electrocatalytic peaks at Hg electrodes [15], including solid amalgam electrodes [20]. Using this electroactive complex it was possible to determine nanomolar and subnanomolar concentrations of the adducts [19]. In the present work, we used streptavidin-modified magnetic beads and biotinylated DNA capture probe (CP) complementary to target miRNA for capturing the miRNA labeled with Os(VI)bipy, followed by EC detection of the labeled adducts. Labeling step is simple and quick due to our newly developed ligand exchange technique [21], in which the sample is first labeled with weakly bound Os(VI)py (py = pyridine) and then the py is replaced with more stable bipy. Whole modification takes no more than 20 min at room temperature. We also show that the protocol has good selectivity and is applicable to real samples by capturing specific miRNA from a total RNA sample containing large number of noncomplementary or similar miRNA sequences. We believe that due to its relative simplicity, short preparation time and low cost the assay is potentially useful in miRNA research.
2. Experimental 2.1. Material Potassium osmate dihydrate, pyridine (py) and 2,2 -bipyridine (bipy) were obtained from Sigma (USA), streptavidin-covered magnetic beads and magnet rack were from Life Technologies (USA), other chemicals were of analytical grade and all aqueous solutions were prepared from deionized water. Biotinylated capture probes
were synthesized by Generi Biotech (Czech Republic), synthetic miRNA-261 and miRNA-522 samples (sequences previously developed to silence recently discovered oncogene AGR2 [23]) were from VBC Biotech GmbH (Austria). Nucleic acids had following sequences: miRNA-261: 5 - AUA UGU CUG AGU CCA GAU GAG -3 cODN-261 (complementary CP to miRNA-261): 5 - ATC TGG ACT CAG ACA TAT TT -3 -biotin miRNA-522: 5 - AUA UCU UCC AGU GAU AUC GGC -3 cODN-522 (complementary CP to miRNA-522): 5 - GAT ATC ACT GGA AGA TAT TT -3 -biotin SM-BIO-522 (single mismatched CP): 5 - GAT ATC ACT GGA AGC TAT TT -3 -biotin SM-MID-522 (single mismatched CP): 5 - GAT ATC ACT TGA AGA TAT TT -3 -biotin DM-522 (double mismatched CP): 5 - GAT ATC ACT TGA AGC TAT TT -3 -biotin In case of miRNA, all nucleotides are ribonucleotides; capture probes contain only deoxyribonucleotides. Underlined letters in CPs represent the mismatch sites when paired with miRNA-522. Sample concentrations were determined spectrophotometrically from absorbance values at 260 nm using NanoDrop ND-1000 spectrophotometer (Thermo Fisher Scientific, USA). 2.2. Labeling process Labeling of miRNA was based on our recently published protocol [21]. Briefly, 10 mM stock Os(VI)py was prepared by suspending 18.4 mg of potassium osmate dihydrate in 4.97 mL water, followed by an addition of 8.9 mL pyridine and 10 mL 10 M HCl. The pH was adjusted to ∼7 and the solution was filtered through 0.45 m Millex membrane (Millipore, USA). Next step was incubation of miRNA sample with 1 mM Os(VI)py for 10 min at 25 ◦ C, followed by an addition of bipy (2 mM) and incubation for another 10 min at 25 ◦ C. This second step led to an exchange of weaker-bound py for bipy, yielding Os(VI)bipy-modified miRNA adducts. In some cases (indicated in the text), unbound Os(VI)bipy label was removed with Amicon Ultra 3 K membrane filters (Millipore, USA). 2.3. Hybridization procedure Magnetic beads covered with streptavidin were pipetted into test tubes and washed three times with following washing buffer: 5 mM Tris-HCl, pH 7.5, 0.5 mM EDTA, 1 M NaCl, followed by incubation with biotinylated capture probe for 15 min at 25 ◦ C. After washing three times, various amounts of miRNA sample was added and incubated for 30 min (at different temperatures). Then, the magnetic beads were thoroughly washed five times to remove any unbound Os(VI)bipy and (in case of a real sample) noncomplementary miRNA sequences. Tubes were heated to 85 ◦ C for 5 min to release labeled miRNA from the beads, and the labeled miRNAs were added to the background electrolyte for EC measurement. 2.4. Preparation of samples from cells Breast cancer-derived cells MCF-7 were cultured in glucoserich Dulbecco’s Modified Eagle’s Medium supplemented with 10% fetal bovine serum at 37 ◦ C in a humidified atmosphere of 5% CO2 . After reaching confluence in 6 cm plate, cells were twice washed by phosphate buffered saline and detached in extraction solution TRI Reagent (MRC, USA) to isolate total cellular RNA. 2.5. Thermal melting curves The formation of nucleic acid duplexes has been studied using MeltDoctor HRM Master Mix (Life Technologies, USA) by melting
M. Bartosik et al. / Analytica Chimica Acta 813 (2014) 35–40
37
Fig. 1. Scheme of the magnetic bead-based assay. Streptavidin-modified magnetic beads (SMBs; 1) are incubated with biotinylated capture probe (2), followed by hybridization of a DNA probe with labeled target miRNA (3), stringent washing to remove unbound label and other interfering molecules (4), thermal release of the miRNA from the magnetic beads (5) and electrochemical measurement (6).
curve analysis on a fast real-time PCR instrument 7900HT (Applied Biosystems, USA). The melting curves were plotted as a first derivative and the inflection point in the melting curve forming a peak was determined as the melting temperature (Tm ) value. 2.6. Instrumentation EC measurements were carried out with Autolab analyzer in connection with VA-Stand 663 (both Metrohm Autolab, Switzerland). The hanging mercury drop electrode (HMDE) was used as a working electrode, Ag/AgCl/3 M KCl as a reference electrode and platinum wire as an auxiliary electrode. All measurements were performed at room temperature using differential pulse voltammetry (DPV). Electrochemical cell containing background electrolyte was deoxygenated prior each measurement by passing a stream of argon for 30 s. 3. Results and discussion In our previous study [16], we have demonstrated that Os(VI)bipy-labeled miRNA of specific sequence can be captured using complementary CP attached to the magnetic beads (MB), followed by the EC determination of the labeled adduct. Although the protocol was relatively inexpensive and quick, several optimization steps were performed in this work to make it even more efficient, bearing in mind that the new protocol should be applicable to real biological samples.
We found several advantages over previously used dT25 -MBs. Whereas before we needed two hybridization events to occur (between dT25 -MBs and adenine tail-containing CP and then between the CP and miRNA), now we could decrease overall assay time with only one hybridization step. Moreover, streptavidinbiotin bond is extremely strong and thus thermal separation of labeled miRNAs prior measurement did not induce CP release. Also, SMBs could be used in experiments when hybridization at elevated temperatures was needed to better discriminate between perfect and mismatched duplexes (see Section 3.6). This would not be possible with dT25 -MBs because elevated temperatures would cause separation of CP from the dT25 -MBs. 3.2. DPV curves of miRNA-522 Fig. 2 shows DP voltammograms of labeled miRNA-522 after the hybridization with its (a) fully complementary capture probe cODN-522 and (b) noncomplementary CP. Os(VI)bipy-miRNA adduct yielded well-developed sharp peak Cat (of electrocatalytic nature) at ∼−1.2 V (vs. Ag/AgCl), while noncomplementary sample displayed no peak at all. Free unbound Os(VI)bipy (sample
3.1. Choice of paramagnetic beads Instead of originally used thymine-modified MBs (dT25 -MBs) designed for capturing nucleic acids containing adenine tail, we have chosen more versatile streptavidin-modified MBs (SMBs). This protocol involved binding of biotinylated CP to SMB, followed by an addition of a sample containing miRNA and subsequent hybridization between the complementary miRNA and CP-modified SMBs (Fig. 1). After the hybridization step, thorough washing was performed to remove any unbound Os(VI)bipy and thus to minimize its contribution to the resulting signal. When working with real samples, this stringent washing also removed other interfering molecules, especially noncomplementary miRNAs. Target miRNA bound to the beads was then thermally released and transferred into the background electrolyte for EC determination.
Fig. 2. DPV curves of labeled miRNA. miRNA-522 with complementary cODN522 (along with the error bar from four independent measurements) shows well-developed electrocatalytic peak Cat (a). Negative controls: miRNA-522 with noncomplementary CP (b) and free Os(VI)bipy reagent without any miRNA (c). Concentrations during hybridization: 100 nM miRNA, 10x excess of free Os(VI)bipy. Background electrolyte: universal Britton-Robinson buffer, pH 4.5 (d). DPV: accumulation time, tA 30 s (with stirring and purging), accumulation potential, EA −0.8 V, initial potential −0.8 V, end potential −1.4 V, step potential 5 mV.
38
M. Bartosik et al. / Analytica Chimica Acta 813 (2014) 35–40
Fig. 3. Dependence of current response on pH of the electrolyte. DPV curves of 500 nM labeled miRNA-522 recorded in the electrolyte with pH 5 (red), pH 4.5 (green), pH 4 (blue) and pH 3.5 (magenta). Other conditions as in Fig. 2 (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.).
without miRNA) was completely washed away, as indicated by an absence of the peak (c). Moreover, unlabeled miRNA also showed no response (not shown), confirming that the peak Cat was due to the Os(VI)bipy-miRNA adducts and not due to unbound label or unlabeled miRNA. 3.3. Reproducibility Peak heights from four independent measurements of the complementary miRNA-522 sample had a relative standard deviation of ∼5% (indicated by an error bar at the peak maximum; Fig. 2). Such high reproducibility is due to an easy, fast and consistent regeneration of the electrode surface, each time producing a new electrode drop with the same atomically smooth surface by simple dislodgment of the old drop. Similar reproducibility of the EC measurements was achieved also for subsequent experiments. In addition to the EC measurement, we have also tested batch-tobatch reproducibility of the whole protocol, i.e. different batches of the miRNA-522 sample underwent hybridization and washing under the same conditions, and found reasonable RSD of ∼10% from four independent assays (not shown). 3.4. Optimization steps Knowing that CP-modified SMBs are suitable for miRNA capture and detection, we then tried to optimize labeling process, hybridization step as well as EC measurement. In our original hybridization procedure, we modified stock solution of the miRNA (usually M) with the Os(VI)bipy and then added a small volume of such labeled miRNA to capture probe-modified MBs, diluting it to desired concentrations. However, if this approach is to be applied in real samples, where miRNA of interest may be available only in small amounts, we needed to test modification efficiency by using different miRNA concentrations during labeling process. We found that lower concentrations of miRNA during labeling step yielded similar peak currents like the original 100 M (not shown), and thus it is not necessary to first modify stock miRNA solution with Os(VI)bipy and then to dilute it for hybridization procedure; rather we can directly modify desired miRNA concentrations (at nM levels) and add them without significant dilution to SMBs. As shown in our previous papers, DPV response from Os(VI)bipy-miRNA is of catalytic nature (i.e., the adduct catalyzes hydrogen evolution at negative potentials), therefore pH of a background electrolyte plays an important role in peak intensity [15,16]. This can be seen in Fig. 3, where a relatively small decrease in pH led to a dramatic increase in current. For instance, at pH 3.5
Fig. 4. Dependence of peak Cat height (iP ) on concentration of Os(VI)bipy-labeled miRNA-522 during hybridization at the SMBs. Error bars from three independent measurements. Other conditions as in Fig. 2.
the peak height is about 4.5-fold larger than at pH 5. This finding was therefore utilized for improving detection limit of our protocol. Although the mechanism of catalysis for the six-valent osmium complexes was not elucidated in detail, it was shown earlier that the DNA modified with eight-valent osmium complex, Os(VIII)bipy, produced similar catalytic signal at ∼−1.2 V (vs. Ag/AgCl) which at acidic pH’s offered highly sensitive determination of DNA [24]. Catalytic peaks at mercury-based electrodes were observed also with other transition metal complexes, including iron [25] and cisplatin [26]. 3.5. Concentration dependence Based on the results from optimization protocol, we performed an experiment in which different concentrations of labeled miRNA522 were added to the SMBs for hybridization with cODN-522. Fig. 4 shows that initial increase in peak Cat height from ∼10 nM to 200 nM concentration was followed by leveling off the signal due to saturation of the electrode surface. At the concentration of ∼10 nM (femtomoles of miRNA-522) the peak height was separated from a blank mean by 3 S.D. Concentrations in the Fig. 4 refer to the values during hybridization, and not necessarily during the measurement. It can be assumed that some miRNA is lost during washing steps and the real concentrations of miRNAs as measured at the HMDE are lower, perhaps in picomolar range. Obviously, efficiency of SMBs to capture as many miRNA molecules present in the sample as possible will be crucial for reaching higher sensitivities. 3.6. Selectivity of the protocol MiRNAs often exhibit high degree of sequence similarity among family members (their sequences may differ only in a single nucleotide), which may pose extra difficulties in their discrimination. Being aware of this complication, we have first designed a set of capture probes which formed either perfect duplex (PD) or single and double mismatched duplexes with the miRNA-522 (SM, DM), followed by an experimental determination of Tm for each duplex in fast real-time PCR system (Fig. 5 Inset). The melting curves revealed that the perfect duplex had the highest Tm of about 57.5 ◦ C, followed by the duplex having a single mismatch at the biotin end (SM-bio; ∼55 ◦ C), duplex with a single mismatch in the middle (SM-MID; ∼48 ◦ C) and finally duplex with a double mismatch (∼42.5 ◦ C). We then hybridized all of the duplexes at the elevated temperature close to the Tm of PD, in order to avoid hybridization of mismatched strands yet keeping the signal from the PD as high as possible. With this approach, we were able to distinguish PD from all other mismatched duplexes (Fig. 5), with the peak heights decreasing in the order PD > SM at the end (43%
M. Bartosik et al. / Analytica Chimica Acta 813 (2014) 35–40
39
Fig. 5. Selectivity protocol. Os-labeled miRNA-522 was hybridized at elevated temperature of 57 ◦ C with complementary capture probe (perfect) and with capture probes having a single mismatch at the end (SM-bio), single mismatch in the middle (SM-MID) and a double mismatch (DM). Peak Cat heights, iP (right), with error bars (from four independent measurements) indicate good discrimination power of the protocol. Inset: Derivative melting curves in duplicate for perfect duplex, SM-bio, SM-MID and DM.
of PD) > SM in the middle (35%) > DM (27%), in agreement with the melting temperature curves. For comparison, noncomplementary sample is shown as well, being negligible. Our results support the evidence that location of a mismatch within the probe sequence is important and that internal mismatches are easier to distinguish than terminal ones [27]. Similar values of discrimination efficiency have been reported also in other EC assays [28,29] and could be possibly improved with the help of various modified probes (LNA, PNA) or by using socalled Y-junctions [11]. Good discrimination efficiency was recently demonstrated also with gold heated electrodes [30]. 3.7. Application to total RNA samples Previous optimization steps led to an improved assay which we decided to test on total RNA samples isolated from in vitro cultured cancer cell lines. We chose the total RNA as a starting material instead of samples containing only miRNAs (purified with commercially available kits) because it not only makes the protocol faster and cheaper, but several orders of magnitude larger background highlights applicability and scalability of our EC approach in real biological samples. We then spiked the total RNA sample (33 ng) with the known concentration of miRNA-261 in various ratios and applied the sample to the SMBs previously modified with the cODN-261 (CP complementary to the miRNA-261). The hybridization occurred at elevated temperature of 57 ◦ C to minimize false positives arising from hybridization of similar sequences. The total RNA sample itself does not contain the miRNA-261 sequence, and thus the negative control experiment, in which no miRNA-261 was added to the total RNA, showed almost no response (Fig. 6). Peak Cat then increased with successive additions of miRNA as expected. We were able to detect target miRNA-261 in four-fold excess of total RNA, which is promising for future applications, but more work will be done to improve this ratio even more. Similar results were obtained also with miRNA-522, although at somewhat lower sensitivity (not shown). In difference to the above experiments (in which any excess label was removed by washing the SMBs), here we added an extra step involving purification of the labeled total RNA samples using Amicon ultracentrifugation 3 kDa filters to completely remove unbound Os(VI)bipy. Without this purification, unbound label would interfere with the measurement, because the total RNA sample was labeled and then used without any significant dilution at the beads, leaving the concentration of the Os(VI)bipy too high to be removed by simple washing of the beads. In the
Fig. 6. Application of the assay to total RNA samples. Column graph shows peak Cat heights (iP ) of target miRNA-261 added at different ratios to total RNA sample (which does not contain miRNA-261) after hybridization at the SMBs at elevated temperature to minimize mismatch contribution. Inset: Zoom of small miRNA-261 concentrations. Other conditions as in Fig. 2.
case of commercial samples, the original millimolar concentration of the Os(VI)bipy was highly diluted and the bead washing sufficed. Although the purification step makes the protocol a bit longer, it might be useful when working with real samples (e.g. cell lysates, body fluids, etc.) containing many small interfering molecules, including nucleotides and mono/oligosaccharides capable of Os(VI)bipy binding, which could be removed by the centrifugation step. 4. Conclusions In this paper we have utilized our previous findings that Os(VI)bipy-labeled miRNA can be detected in a very simple, inexpensive and fast assay using magnetic beads and HMDE [16] by optimizing various assay steps that led to an improved sensitivity, selectivity and applicability to total RNA samples. The labeling step is easy and the resulting adducts yield sensitive peaks at the HDME due to their electrocatalytic nature, where single Os(VI)bipymiRNA adduct may catalyze reduction of multiple hydrogen ions to generate molecular hydrogen, leading to an increased signal. No other amplification step was needed, although it is not excluded that in the future some kind of amplification will be desirable in order to achieve lower detection limits. One option is to apply antibodies, widely used in immunosensing. Indeed, Os(VIII)bipymodified nucleic acids were highly immunogenic, making possible
40
M. Bartosik et al. / Analytica Chimica Acta 813 (2014) 35–40
easy generation of poly- and monoclonal antibodies [31]. Moreover, antibodies against Os(VI)L-RNA adducts have been recently obtained, which could be applied in the analysis of miRNA. An important and often overlooked aspect of the miRNA detection is discrimination between related miRNA sequences, leading to false positives if proper care is not taken. Here we address this issue using a simple approach to discriminate between perfect duplex and single/double mismatches. This is especially important when working with real samples in which high excess of highly similar miRNA sequences can be present. We have demonstrated feasibility of our assay by detecting target miRNA in higher concentration of total RNA samples. We hope that our newly developed assay will be applicable for detection of both miRNA aberrantly expressed in human cancer cells and of circulating miRNA found in body fluids. Acknowledgements This work was supported by grants from GACR P206/12/G151, GACR P301/13/00956S, the European Regional Development Fund and the State Budget of the Czech Republic (RECAMO CZ.1.05/2.1.00/03.0101) and by MH CZ-DRO (MMCI, 00209805). References [1] M.V. Iorio, C.M. Croce, Carcinogenesis 33 (2012) 1126–1133. [2] H.F. Dong, J.P. Lei, L. Ding, Y.Q. Wen, H.X. Ju, X.J. Zhang, Chem. Rev. 113 (2013) 6207–6233. [3] J. Liu, S.F. Jennings, W. Tong, H. Hong, J. Biomed. Sci. Eng. 4 (2011) 666–676. [4] K.A. Cissell, S.K. Deo, Anal. Bioanal. Chem. 394 (2009) 1109–1116. [5] M. de Planell-Saguer, M.C. Rodicio, Anal. Chim. Acta 699 (2011) 134–152.
[6] Z.Q. Gao, H.M. Deng, W. Shen, Y.Q. Ren, Anal. Chem. 85 (2013) 1624–1630. [7] Y.Q. Ren, H.M. Deng, W. Shen, Z.Q. Gao, Anal. Chem. 85 (2013) 4784–4789. [8] H.V. Tran, B. Piro, S. Reisberg, L.D. Tran, H.T. Duc, M.C. Pham, Biosens. Bioelectron. 49 (2013) 164–169. [9] H.S. Yin, Y.L. Zhou, H.X. Zhang, X.M. Meng, S.Y. Ai, Biosens. Bioelectron. 33 (2012) 247–253. [10] T. Kilic, S.N. Topkaya, D.O. Ariksoysal, M. Ozsoz, P. Ballar, Y. Erac, O. Gozen, Biosens. Bioelectron. 38 (2012) 195–201. [11] Z.W. Wang, J. Zhang, Y. Guo, X.Y. Wu, W.J. Yang, L.J. Xu, J.H. Chen, F.F. Fu, Biosens. Bioelectron. 45 (2013) 108–113. [12] F. Bettazzi, E. Hamid-Asl, C.L. Esposito, C. Quintavalle, N. Formisano, S. Laschi, S. Catuogno, M. Iaboni, G. Marrazza, M. Mascini, L. Cerchia, V. De Franciscis, G. Condorelli, I. Palchetti, Anal. Bioanal. Chem. 405 (2013) 1025–1034. [13] J. Wang, X. Yi, H. Tang, H. Han, M. Wu, F. Zhou, Anal. Chem. 84 (2012) 6400–6406. [14] Z.Q. Gao, Y.H. Yu, Biosens. Bioelectron. 22 (2007) 933–940. [15] M. Trefulka, M. Bartosik, E. Palecek, Electrochem. Commun. 12 (2010) 1760–1763. [16] M. Bartosik, M. Trefulka, R. Hrstka, B. Vojtesek, E. Palecek, Electrochem. Commun. 33 (2013) 55–58. [17] E. Palecek, Meth. Enzymol. 212 (1992) 139–155. [18] M. Trefulka, E. Palecek, Electroanalysis 22 (2010) 1837–1845. [19] M. Trefulka, E. Palecek, Bioelectrochemistry 88 (2012) 8–14. [20] E. Palecek, M. Trefulka, Analyst 136 (2011) 321–326. [21] M. Trefulka, E. Palecek, Electroanalysis 25 (2013) 1813–1817. [22] M. Trefulka, V. Ostatna, L. Havran, M. Fojta, E. Palecek, Electroanalysis 19 (2007) 1281–1287. [23] R. Hrstka, R. Nenutil, A. Fourtouna, M.M. Maslon, C. Naughton, S. Langdon, E. Murray, A. Larionov, K. Petrakova, P. Muller, M.J. Dixon, T.R. Hupp, B. Vojtesek, Oncogene 29 (2010) 4838–4847. [24] R. Kizek, L. Havran, M. Fojta, E. Palecek, Bioelectrochemistry 55 (2002) 119–121. [25] M. Heyrovsky, Collect. Czech. Chem. Commun. 66 (2001) 67–80. [26] P. Horakova, L. Tesnohlidkova, L. Havran, P. Vidlakova, H. Pivonkova, M. Fojta, Anal. Chem. 82 (2010) 2969–2976. [27] Y. You, B.G. Moreira, M.A. Behlke, R. Owczarzy, Nucl. Acids Res. 34 (2006) e60. [28] H. Yang, A. Hui, G. Pampalakis, L. Soleymani, F.F. Liu, E.H. Sargent, S.O. Kelley, Angew. Chem. Int. Ed. 48 (2009) 8461–8464. [29] C. Pohlmann, M. Sprinzl, Anal. Chem. 82 (2010) 4434–4440. [30] M. Jacobsen, G.U. Flechsig, Electroanalysis 25 (2013) 373–379. [31] E. Palecek, M. Bartosik, Chem. Rev. 112 (2012) 3427–3481.
Contents lists available at ScienceDirect
Analytica Chimica Acta journal homepage: www.elsevier.com/locate/aca
Magnetic bead-based hybridization assay for electrochemical detection of microRNA Martin Bartosik a,b,∗ , Roman Hrstka a , Emil Palecek a,b , Borivoj Vojtesek a a b
Regional Centre for Applied Molecular Oncology (RECAMO), Masaryk Memorial Cancer Institute, Zluty kopec 7, 656 53 Brno, Czech Republic Institute of Biophysics, Academy of Sciences of the Czech Republic, v.v.i., Kralovopolska 135, 612 65 Brno, Czech Republic
h i g h l i g h t s
g r a p h i c a l
a b s t r a c t
• Simple and quick electrochemical assay for microRNA detection was developed. • Specific electroactive complex based on osmium(VI) was used for microRNA 3 -end labeling. • Good selectivity was obtained by discriminating perfect duplex and single mismatches. • Labeled microRNA could be detected in total RNA samples.
a r t i c l e
i n f o
Article history: Received 12 November 2013 Received in revised form 6 January 2014 Accepted 8 January 2014 Available online 16 January 2014 Keywords: MicroRNA Electrochemistry Mercury electrodes Mismatch discrimination
a b s t r a c t Aberrant expression of microRNAs (miRNAs), short non-coding RNA molecules regulating gene expression, is often found in tumor cells, making the miRNAs suitable candidates as cancer biomarkers. Electrochemistry is an interesting alternative to current standard methods of miRNA detection by offering cheaper instrumentation and faster assays times. In this paper, we labeled miRNA in a quick, simple, two-step procedure with electroactive complex of osmium(VI) and 2,2 -bipyridine, Os(VI)bipy, which specifically binds to the ribose at the 3 -end of the miRNA, and hybridized such labeled miRNA with biotinylated capture probe attached to the streptavidin magnetic beads. Labeled miRNA was then detected at hanging mercury drop electrode at femtomole level due to an electrocatalytic nature of the peak from the Os(VI)bipy label. We obtained good selectivity of the assay using elevated hybridization temperatures for better discrimination of perfect duplex from single and double mismatches. After optimization of the protocol, we demonstrated feasibility of our assay by detecting target miRNA in real total RNA samples isolated from human cancer cells. © 2014 Elsevier B.V. All rights reserved.
1. Introduction A boom in microRNA (miRNA) research can be observed in past few years, especially due to the finding that altered expression levels of miRNAs in cells are associated with onset and progress of the cancer, making them attractive not only as diagnostic biomarkers, but also as potential therapeutic targets [1]. The link between miRNAs and many diseases is due to their binding to mRNA and
∗ Corresponding author at: RECAMO. Tel.: +420 543 133 325. E-mail address: [email protected] (M. Bartosik). 0003-2670/$ – see front matter © 2014 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.aca.2014.01.023
subsequent negative regulation of gene expression, including many oncogenes and tumor suppressor genes. Most-widely employed method for detection of aberrantly expressed miRNAs is qRT-PCR, where the reverse transcription (RT) first converts target miRNA into the cDNA, followed by its amplification using specifically designed primers and by quantification of the amplified duplexes [2]. The method has a good accuracy and sensitivity, enabling determination of low levels of miRNAs, but the amplification step is not straightforward because miRNAs are usually much shorter than primers. Later, microarrays were introduced for high-throughput and rapid multiplexed detection of miRNAs, both with and without PCR amplification. They are based
36
M. Bartosik et al. / Analytica Chimica Acta 813 (2014) 35–40
on nucleic acid hybridization between fluorescently labeled miRNAs and solid substrate-attached capture probes, followed by a fluorescence read-out of the label. Recently, profiling of miRNAs by next-generation sequencing (NGS) has progressed rapidly. Several NGS platforms were established allowing not only measurement of expression levels of known miRNAs, but also detection of unknown miRNAs [3]. These robust techniques, however, bear higher costs due to sophisticated instruments and demand for highly skilled personnel. Therefore, new methods for miRNA detection are being developed which would be fast, inexpensive and relatively simple [4,5]. Such requirements are met by electrochemical (EC) techniques, which besides above-mentioned advantages also offer possibility for miniaturization and parallel measurements of multiple samples at electrode chips. More and more papers appear each year also for EC-based detection of miRNAs, both in label-free [6–8] and label-based format [9–14]. The latter one, although more complex, usually provides higher sensitivity and versatility by introducing e.g. specific antibodies (conjugated with enzymes and detecting enzymatic products) [9–12], or nanoparticle tags [13,14]. Recently, we have used an electroactive label based on sixvalent osmium, Os(VI), in complex with nitrogenous ligand (L), Os(VI)L, for specific modification of 3 -terminal ribose in RNA [15,16]. This complex is similar to the eight-valent Os(VIII)L, which was developed already in 1980’s for probing the DNA structure by binding to unpaired pyrimidine residues in single-stranded portions of DNA [17]. Labeling of miRNAs with these Os(VIII) complexes would be possible but modification of bases (uracil and cytosine) would interfere with the miRNA hybridization necessary for determination of specific miRNAs. This is not the case with Os(VI)L complexes which were shown to react specifically with diol groups in sugar residues in oligo- and polysaccharides [18–21] as well as in ribosides [22], but not in deoxyribosides (lacking the diol group at the 3 -end of the molecule) or with nucleic acid bases. We mostly relied on 2,2 -bipyridine (bipy) as a ligand for the Os(VI) because Os(VI)bipy-modified adducts yield redox couples at carbon and highly sensitive electrocatalytic peaks at Hg electrodes [15], including solid amalgam electrodes [20]. Using this electroactive complex it was possible to determine nanomolar and subnanomolar concentrations of the adducts [19]. In the present work, we used streptavidin-modified magnetic beads and biotinylated DNA capture probe (CP) complementary to target miRNA for capturing the miRNA labeled with Os(VI)bipy, followed by EC detection of the labeled adducts. Labeling step is simple and quick due to our newly developed ligand exchange technique [21], in which the sample is first labeled with weakly bound Os(VI)py (py = pyridine) and then the py is replaced with more stable bipy. Whole modification takes no more than 20 min at room temperature. We also show that the protocol has good selectivity and is applicable to real samples by capturing specific miRNA from a total RNA sample containing large number of noncomplementary or similar miRNA sequences. We believe that due to its relative simplicity, short preparation time and low cost the assay is potentially useful in miRNA research.
2. Experimental 2.1. Material Potassium osmate dihydrate, pyridine (py) and 2,2 -bipyridine (bipy) were obtained from Sigma (USA), streptavidin-covered magnetic beads and magnet rack were from Life Technologies (USA), other chemicals were of analytical grade and all aqueous solutions were prepared from deionized water. Biotinylated capture probes
were synthesized by Generi Biotech (Czech Republic), synthetic miRNA-261 and miRNA-522 samples (sequences previously developed to silence recently discovered oncogene AGR2 [23]) were from VBC Biotech GmbH (Austria). Nucleic acids had following sequences: miRNA-261: 5 - AUA UGU CUG AGU CCA GAU GAG -3 cODN-261 (complementary CP to miRNA-261): 5 - ATC TGG ACT CAG ACA TAT TT -3 -biotin miRNA-522: 5 - AUA UCU UCC AGU GAU AUC GGC -3 cODN-522 (complementary CP to miRNA-522): 5 - GAT ATC ACT GGA AGA TAT TT -3 -biotin SM-BIO-522 (single mismatched CP): 5 - GAT ATC ACT GGA AGC TAT TT -3 -biotin SM-MID-522 (single mismatched CP): 5 - GAT ATC ACT TGA AGA TAT TT -3 -biotin DM-522 (double mismatched CP): 5 - GAT ATC ACT TGA AGC TAT TT -3 -biotin In case of miRNA, all nucleotides are ribonucleotides; capture probes contain only deoxyribonucleotides. Underlined letters in CPs represent the mismatch sites when paired with miRNA-522. Sample concentrations were determined spectrophotometrically from absorbance values at 260 nm using NanoDrop ND-1000 spectrophotometer (Thermo Fisher Scientific, USA). 2.2. Labeling process Labeling of miRNA was based on our recently published protocol [21]. Briefly, 10 mM stock Os(VI)py was prepared by suspending 18.4 mg of potassium osmate dihydrate in 4.97 mL water, followed by an addition of 8.9 mL pyridine and 10 mL 10 M HCl. The pH was adjusted to ∼7 and the solution was filtered through 0.45 m Millex membrane (Millipore, USA). Next step was incubation of miRNA sample with 1 mM Os(VI)py for 10 min at 25 ◦ C, followed by an addition of bipy (2 mM) and incubation for another 10 min at 25 ◦ C. This second step led to an exchange of weaker-bound py for bipy, yielding Os(VI)bipy-modified miRNA adducts. In some cases (indicated in the text), unbound Os(VI)bipy label was removed with Amicon Ultra 3 K membrane filters (Millipore, USA). 2.3. Hybridization procedure Magnetic beads covered with streptavidin were pipetted into test tubes and washed three times with following washing buffer: 5 mM Tris-HCl, pH 7.5, 0.5 mM EDTA, 1 M NaCl, followed by incubation with biotinylated capture probe for 15 min at 25 ◦ C. After washing three times, various amounts of miRNA sample was added and incubated for 30 min (at different temperatures). Then, the magnetic beads were thoroughly washed five times to remove any unbound Os(VI)bipy and (in case of a real sample) noncomplementary miRNA sequences. Tubes were heated to 85 ◦ C for 5 min to release labeled miRNA from the beads, and the labeled miRNAs were added to the background electrolyte for EC measurement. 2.4. Preparation of samples from cells Breast cancer-derived cells MCF-7 were cultured in glucoserich Dulbecco’s Modified Eagle’s Medium supplemented with 10% fetal bovine serum at 37 ◦ C in a humidified atmosphere of 5% CO2 . After reaching confluence in 6 cm plate, cells were twice washed by phosphate buffered saline and detached in extraction solution TRI Reagent (MRC, USA) to isolate total cellular RNA. 2.5. Thermal melting curves The formation of nucleic acid duplexes has been studied using MeltDoctor HRM Master Mix (Life Technologies, USA) by melting
M. Bartosik et al. / Analytica Chimica Acta 813 (2014) 35–40
37
Fig. 1. Scheme of the magnetic bead-based assay. Streptavidin-modified magnetic beads (SMBs; 1) are incubated with biotinylated capture probe (2), followed by hybridization of a DNA probe with labeled target miRNA (3), stringent washing to remove unbound label and other interfering molecules (4), thermal release of the miRNA from the magnetic beads (5) and electrochemical measurement (6).
curve analysis on a fast real-time PCR instrument 7900HT (Applied Biosystems, USA). The melting curves were plotted as a first derivative and the inflection point in the melting curve forming a peak was determined as the melting temperature (Tm ) value. 2.6. Instrumentation EC measurements were carried out with Autolab analyzer in connection with VA-Stand 663 (both Metrohm Autolab, Switzerland). The hanging mercury drop electrode (HMDE) was used as a working electrode, Ag/AgCl/3 M KCl as a reference electrode and platinum wire as an auxiliary electrode. All measurements were performed at room temperature using differential pulse voltammetry (DPV). Electrochemical cell containing background electrolyte was deoxygenated prior each measurement by passing a stream of argon for 30 s. 3. Results and discussion In our previous study [16], we have demonstrated that Os(VI)bipy-labeled miRNA of specific sequence can be captured using complementary CP attached to the magnetic beads (MB), followed by the EC determination of the labeled adduct. Although the protocol was relatively inexpensive and quick, several optimization steps were performed in this work to make it even more efficient, bearing in mind that the new protocol should be applicable to real biological samples.
We found several advantages over previously used dT25 -MBs. Whereas before we needed two hybridization events to occur (between dT25 -MBs and adenine tail-containing CP and then between the CP and miRNA), now we could decrease overall assay time with only one hybridization step. Moreover, streptavidinbiotin bond is extremely strong and thus thermal separation of labeled miRNAs prior measurement did not induce CP release. Also, SMBs could be used in experiments when hybridization at elevated temperatures was needed to better discriminate between perfect and mismatched duplexes (see Section 3.6). This would not be possible with dT25 -MBs because elevated temperatures would cause separation of CP from the dT25 -MBs. 3.2. DPV curves of miRNA-522 Fig. 2 shows DP voltammograms of labeled miRNA-522 after the hybridization with its (a) fully complementary capture probe cODN-522 and (b) noncomplementary CP. Os(VI)bipy-miRNA adduct yielded well-developed sharp peak Cat (of electrocatalytic nature) at ∼−1.2 V (vs. Ag/AgCl), while noncomplementary sample displayed no peak at all. Free unbound Os(VI)bipy (sample
3.1. Choice of paramagnetic beads Instead of originally used thymine-modified MBs (dT25 -MBs) designed for capturing nucleic acids containing adenine tail, we have chosen more versatile streptavidin-modified MBs (SMBs). This protocol involved binding of biotinylated CP to SMB, followed by an addition of a sample containing miRNA and subsequent hybridization between the complementary miRNA and CP-modified SMBs (Fig. 1). After the hybridization step, thorough washing was performed to remove any unbound Os(VI)bipy and thus to minimize its contribution to the resulting signal. When working with real samples, this stringent washing also removed other interfering molecules, especially noncomplementary miRNAs. Target miRNA bound to the beads was then thermally released and transferred into the background electrolyte for EC determination.
Fig. 2. DPV curves of labeled miRNA. miRNA-522 with complementary cODN522 (along with the error bar from four independent measurements) shows well-developed electrocatalytic peak Cat (a). Negative controls: miRNA-522 with noncomplementary CP (b) and free Os(VI)bipy reagent without any miRNA (c). Concentrations during hybridization: 100 nM miRNA, 10x excess of free Os(VI)bipy. Background electrolyte: universal Britton-Robinson buffer, pH 4.5 (d). DPV: accumulation time, tA 30 s (with stirring and purging), accumulation potential, EA −0.8 V, initial potential −0.8 V, end potential −1.4 V, step potential 5 mV.
38
M. Bartosik et al. / Analytica Chimica Acta 813 (2014) 35–40
Fig. 3. Dependence of current response on pH of the electrolyte. DPV curves of 500 nM labeled miRNA-522 recorded in the electrolyte with pH 5 (red), pH 4.5 (green), pH 4 (blue) and pH 3.5 (magenta). Other conditions as in Fig. 2 (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.).
without miRNA) was completely washed away, as indicated by an absence of the peak (c). Moreover, unlabeled miRNA also showed no response (not shown), confirming that the peak Cat was due to the Os(VI)bipy-miRNA adducts and not due to unbound label or unlabeled miRNA. 3.3. Reproducibility Peak heights from four independent measurements of the complementary miRNA-522 sample had a relative standard deviation of ∼5% (indicated by an error bar at the peak maximum; Fig. 2). Such high reproducibility is due to an easy, fast and consistent regeneration of the electrode surface, each time producing a new electrode drop with the same atomically smooth surface by simple dislodgment of the old drop. Similar reproducibility of the EC measurements was achieved also for subsequent experiments. In addition to the EC measurement, we have also tested batch-tobatch reproducibility of the whole protocol, i.e. different batches of the miRNA-522 sample underwent hybridization and washing under the same conditions, and found reasonable RSD of ∼10% from four independent assays (not shown). 3.4. Optimization steps Knowing that CP-modified SMBs are suitable for miRNA capture and detection, we then tried to optimize labeling process, hybridization step as well as EC measurement. In our original hybridization procedure, we modified stock solution of the miRNA (usually M) with the Os(VI)bipy and then added a small volume of such labeled miRNA to capture probe-modified MBs, diluting it to desired concentrations. However, if this approach is to be applied in real samples, where miRNA of interest may be available only in small amounts, we needed to test modification efficiency by using different miRNA concentrations during labeling process. We found that lower concentrations of miRNA during labeling step yielded similar peak currents like the original 100 M (not shown), and thus it is not necessary to first modify stock miRNA solution with Os(VI)bipy and then to dilute it for hybridization procedure; rather we can directly modify desired miRNA concentrations (at nM levels) and add them without significant dilution to SMBs. As shown in our previous papers, DPV response from Os(VI)bipy-miRNA is of catalytic nature (i.e., the adduct catalyzes hydrogen evolution at negative potentials), therefore pH of a background electrolyte plays an important role in peak intensity [15,16]. This can be seen in Fig. 3, where a relatively small decrease in pH led to a dramatic increase in current. For instance, at pH 3.5
Fig. 4. Dependence of peak Cat height (iP ) on concentration of Os(VI)bipy-labeled miRNA-522 during hybridization at the SMBs. Error bars from three independent measurements. Other conditions as in Fig. 2.
the peak height is about 4.5-fold larger than at pH 5. This finding was therefore utilized for improving detection limit of our protocol. Although the mechanism of catalysis for the six-valent osmium complexes was not elucidated in detail, it was shown earlier that the DNA modified with eight-valent osmium complex, Os(VIII)bipy, produced similar catalytic signal at ∼−1.2 V (vs. Ag/AgCl) which at acidic pH’s offered highly sensitive determination of DNA [24]. Catalytic peaks at mercury-based electrodes were observed also with other transition metal complexes, including iron [25] and cisplatin [26]. 3.5. Concentration dependence Based on the results from optimization protocol, we performed an experiment in which different concentrations of labeled miRNA522 were added to the SMBs for hybridization with cODN-522. Fig. 4 shows that initial increase in peak Cat height from ∼10 nM to 200 nM concentration was followed by leveling off the signal due to saturation of the electrode surface. At the concentration of ∼10 nM (femtomoles of miRNA-522) the peak height was separated from a blank mean by 3 S.D. Concentrations in the Fig. 4 refer to the values during hybridization, and not necessarily during the measurement. It can be assumed that some miRNA is lost during washing steps and the real concentrations of miRNAs as measured at the HMDE are lower, perhaps in picomolar range. Obviously, efficiency of SMBs to capture as many miRNA molecules present in the sample as possible will be crucial for reaching higher sensitivities. 3.6. Selectivity of the protocol MiRNAs often exhibit high degree of sequence similarity among family members (their sequences may differ only in a single nucleotide), which may pose extra difficulties in their discrimination. Being aware of this complication, we have first designed a set of capture probes which formed either perfect duplex (PD) or single and double mismatched duplexes with the miRNA-522 (SM, DM), followed by an experimental determination of Tm for each duplex in fast real-time PCR system (Fig. 5 Inset). The melting curves revealed that the perfect duplex had the highest Tm of about 57.5 ◦ C, followed by the duplex having a single mismatch at the biotin end (SM-bio; ∼55 ◦ C), duplex with a single mismatch in the middle (SM-MID; ∼48 ◦ C) and finally duplex with a double mismatch (∼42.5 ◦ C). We then hybridized all of the duplexes at the elevated temperature close to the Tm of PD, in order to avoid hybridization of mismatched strands yet keeping the signal from the PD as high as possible. With this approach, we were able to distinguish PD from all other mismatched duplexes (Fig. 5), with the peak heights decreasing in the order PD > SM at the end (43%
M. Bartosik et al. / Analytica Chimica Acta 813 (2014) 35–40
39
Fig. 5. Selectivity protocol. Os-labeled miRNA-522 was hybridized at elevated temperature of 57 ◦ C with complementary capture probe (perfect) and with capture probes having a single mismatch at the end (SM-bio), single mismatch in the middle (SM-MID) and a double mismatch (DM). Peak Cat heights, iP (right), with error bars (from four independent measurements) indicate good discrimination power of the protocol. Inset: Derivative melting curves in duplicate for perfect duplex, SM-bio, SM-MID and DM.
of PD) > SM in the middle (35%) > DM (27%), in agreement with the melting temperature curves. For comparison, noncomplementary sample is shown as well, being negligible. Our results support the evidence that location of a mismatch within the probe sequence is important and that internal mismatches are easier to distinguish than terminal ones [27]. Similar values of discrimination efficiency have been reported also in other EC assays [28,29] and could be possibly improved with the help of various modified probes (LNA, PNA) or by using socalled Y-junctions [11]. Good discrimination efficiency was recently demonstrated also with gold heated electrodes [30]. 3.7. Application to total RNA samples Previous optimization steps led to an improved assay which we decided to test on total RNA samples isolated from in vitro cultured cancer cell lines. We chose the total RNA as a starting material instead of samples containing only miRNAs (purified with commercially available kits) because it not only makes the protocol faster and cheaper, but several orders of magnitude larger background highlights applicability and scalability of our EC approach in real biological samples. We then spiked the total RNA sample (33 ng) with the known concentration of miRNA-261 in various ratios and applied the sample to the SMBs previously modified with the cODN-261 (CP complementary to the miRNA-261). The hybridization occurred at elevated temperature of 57 ◦ C to minimize false positives arising from hybridization of similar sequences. The total RNA sample itself does not contain the miRNA-261 sequence, and thus the negative control experiment, in which no miRNA-261 was added to the total RNA, showed almost no response (Fig. 6). Peak Cat then increased with successive additions of miRNA as expected. We were able to detect target miRNA-261 in four-fold excess of total RNA, which is promising for future applications, but more work will be done to improve this ratio even more. Similar results were obtained also with miRNA-522, although at somewhat lower sensitivity (not shown). In difference to the above experiments (in which any excess label was removed by washing the SMBs), here we added an extra step involving purification of the labeled total RNA samples using Amicon ultracentrifugation 3 kDa filters to completely remove unbound Os(VI)bipy. Without this purification, unbound label would interfere with the measurement, because the total RNA sample was labeled and then used without any significant dilution at the beads, leaving the concentration of the Os(VI)bipy too high to be removed by simple washing of the beads. In the
Fig. 6. Application of the assay to total RNA samples. Column graph shows peak Cat heights (iP ) of target miRNA-261 added at different ratios to total RNA sample (which does not contain miRNA-261) after hybridization at the SMBs at elevated temperature to minimize mismatch contribution. Inset: Zoom of small miRNA-261 concentrations. Other conditions as in Fig. 2.
case of commercial samples, the original millimolar concentration of the Os(VI)bipy was highly diluted and the bead washing sufficed. Although the purification step makes the protocol a bit longer, it might be useful when working with real samples (e.g. cell lysates, body fluids, etc.) containing many small interfering molecules, including nucleotides and mono/oligosaccharides capable of Os(VI)bipy binding, which could be removed by the centrifugation step. 4. Conclusions In this paper we have utilized our previous findings that Os(VI)bipy-labeled miRNA can be detected in a very simple, inexpensive and fast assay using magnetic beads and HMDE [16] by optimizing various assay steps that led to an improved sensitivity, selectivity and applicability to total RNA samples. The labeling step is easy and the resulting adducts yield sensitive peaks at the HDME due to their electrocatalytic nature, where single Os(VI)bipymiRNA adduct may catalyze reduction of multiple hydrogen ions to generate molecular hydrogen, leading to an increased signal. No other amplification step was needed, although it is not excluded that in the future some kind of amplification will be desirable in order to achieve lower detection limits. One option is to apply antibodies, widely used in immunosensing. Indeed, Os(VIII)bipymodified nucleic acids were highly immunogenic, making possible
40
M. Bartosik et al. / Analytica Chimica Acta 813 (2014) 35–40
easy generation of poly- and monoclonal antibodies [31]. Moreover, antibodies against Os(VI)L-RNA adducts have been recently obtained, which could be applied in the analysis of miRNA. An important and often overlooked aspect of the miRNA detection is discrimination between related miRNA sequences, leading to false positives if proper care is not taken. Here we address this issue using a simple approach to discriminate between perfect duplex and single/double mismatches. This is especially important when working with real samples in which high excess of highly similar miRNA sequences can be present. We have demonstrated feasibility of our assay by detecting target miRNA in higher concentration of total RNA samples. We hope that our newly developed assay will be applicable for detection of both miRNA aberrantly expressed in human cancer cells and of circulating miRNA found in body fluids. Acknowledgements This work was supported by grants from GACR P206/12/G151, GACR P301/13/00956S, the European Regional Development Fund and the State Budget of the Czech Republic (RECAMO CZ.1.05/2.1.00/03.0101) and by MH CZ-DRO (MMCI, 00209805). References [1] M.V. Iorio, C.M. Croce, Carcinogenesis 33 (2012) 1126–1133. [2] H.F. Dong, J.P. Lei, L. Ding, Y.Q. Wen, H.X. Ju, X.J. Zhang, Chem. Rev. 113 (2013) 6207–6233. [3] J. Liu, S.F. Jennings, W. Tong, H. Hong, J. Biomed. Sci. Eng. 4 (2011) 666–676. [4] K.A. Cissell, S.K. Deo, Anal. Bioanal. Chem. 394 (2009) 1109–1116. [5] M. de Planell-Saguer, M.C. Rodicio, Anal. Chim. Acta 699 (2011) 134–152.
[6] Z.Q. Gao, H.M. Deng, W. Shen, Y.Q. Ren, Anal. Chem. 85 (2013) 1624–1630. [7] Y.Q. Ren, H.M. Deng, W. Shen, Z.Q. Gao, Anal. Chem. 85 (2013) 4784–4789. [8] H.V. Tran, B. Piro, S. Reisberg, L.D. Tran, H.T. Duc, M.C. Pham, Biosens. Bioelectron. 49 (2013) 164–169. [9] H.S. Yin, Y.L. Zhou, H.X. Zhang, X.M. Meng, S.Y. Ai, Biosens. Bioelectron. 33 (2012) 247–253. [10] T. Kilic, S.N. Topkaya, D.O. Ariksoysal, M. Ozsoz, P. Ballar, Y. Erac, O. Gozen, Biosens. Bioelectron. 38 (2012) 195–201. [11] Z.W. Wang, J. Zhang, Y. Guo, X.Y. Wu, W.J. Yang, L.J. Xu, J.H. Chen, F.F. Fu, Biosens. Bioelectron. 45 (2013) 108–113. [12] F. Bettazzi, E. Hamid-Asl, C.L. Esposito, C. Quintavalle, N. Formisano, S. Laschi, S. Catuogno, M. Iaboni, G. Marrazza, M. Mascini, L. Cerchia, V. De Franciscis, G. Condorelli, I. Palchetti, Anal. Bioanal. Chem. 405 (2013) 1025–1034. [13] J. Wang, X. Yi, H. Tang, H. Han, M. Wu, F. Zhou, Anal. Chem. 84 (2012) 6400–6406. [14] Z.Q. Gao, Y.H. Yu, Biosens. Bioelectron. 22 (2007) 933–940. [15] M. Trefulka, M. Bartosik, E. Palecek, Electrochem. Commun. 12 (2010) 1760–1763. [16] M. Bartosik, M. Trefulka, R. Hrstka, B. Vojtesek, E. Palecek, Electrochem. Commun. 33 (2013) 55–58. [17] E. Palecek, Meth. Enzymol. 212 (1992) 139–155. [18] M. Trefulka, E. Palecek, Electroanalysis 22 (2010) 1837–1845. [19] M. Trefulka, E. Palecek, Bioelectrochemistry 88 (2012) 8–14. [20] E. Palecek, M. Trefulka, Analyst 136 (2011) 321–326. [21] M. Trefulka, E. Palecek, Electroanalysis 25 (2013) 1813–1817. [22] M. Trefulka, V. Ostatna, L. Havran, M. Fojta, E. Palecek, Electroanalysis 19 (2007) 1281–1287. [23] R. Hrstka, R. Nenutil, A. Fourtouna, M.M. Maslon, C. Naughton, S. Langdon, E. Murray, A. Larionov, K. Petrakova, P. Muller, M.J. Dixon, T.R. Hupp, B. Vojtesek, Oncogene 29 (2010) 4838–4847. [24] R. Kizek, L. Havran, M. Fojta, E. Palecek, Bioelectrochemistry 55 (2002) 119–121. [25] M. Heyrovsky, Collect. Czech. Chem. Commun. 66 (2001) 67–80. [26] P. Horakova, L. Tesnohlidkova, L. Havran, P. Vidlakova, H. Pivonkova, M. Fojta, Anal. Chem. 82 (2010) 2969–2976. [27] Y. You, B.G. Moreira, M.A. Behlke, R. Owczarzy, Nucl. Acids Res. 34 (2006) e60. [28] H. Yang, A. Hui, G. Pampalakis, L. Soleymani, F.F. Liu, E.H. Sargent, S.O. Kelley, Angew. Chem. Int. Ed. 48 (2009) 8461–8464. [29] C. Pohlmann, M. Sprinzl, Anal. Chem. 82 (2010) 4434–4440. [30] M. Jacobsen, G.U. Flechsig, Electroanalysis 25 (2013) 373–379. [31] E. Palecek, M. Bartosik, Chem. Rev. 112 (2012) 3427–3481.
