Electrochemical Labeling of DNA Hybridization Probes with Osmium

Dec 22, 2006 - electroactive markers has frequently been utilized in electrochemical detection of DNA hybridization. Osmium tetroxide complexes with t...
0 downloads 0 Views 379KB Size
Download PDF
Anal. Chem. 2007, 79, 1022-1029

“Multicolor” Electrochemical Labeling of DNA Hybridization Probes with Osmium Tetroxide Complexes Miroslav Fojta,* Pavel Kostecˇka, Mojmı´r Trefulka, Ludeˇk Havran, and Emil Palecˇek

Institute of Biophysics, Academy of Sciences of the Czech Republic, Kralovopolska 135, 612 65 Brno, Czech Republic

Labeling of oligonucleotide reporter probes (RP) with electroactive markers has frequently been utilized in electrochemical detection of DNA hybridization. Osmium tetroxide complexes with tertiary amines (Os,L) bind covalently to pyrimidine (predominantly thymine) bases in DNA, forming stable, electrochemically active adducts. We propose a technique of electrochemical “multicolor” DNA coding based on RP labeling with Os,L markers involving different nitrogenous ligands (such as 2,2′ -bipyridine, 1,10-phenanthroline derivatives or N,N,N′,N′tetramethylethylenediamine). At carbon electrodes the Os,L-labeled RPs produce specific signals, with the potentials of these differing depending on the ligand type. When using Os,L markers providing sufficiently large differences in their peak potentials, parallel analysis of multiple target DNA sequences can easily be performed via DNA hybridization at magnetic beads followed by voltammetric detection at carbon electrodes. Os,L labeling of oligonucleotide probes comprising a segment complementary to target DNA and an oligo(T) tail (to be modified with the osmium complex) does not require any organic chemistry facilities and can be achieved in any molecular biological laboratory. We also for the first time show that this technology can be used for labeling of oligonucleotide probes hybridizing with target DNAs that contain both purine and pyrimidine bases.

metal species has proved particularly convenient due to wellpronounced electrochemical activity of the metals. Organic chelates of transition metal such as ferrocene, ruthenium, or osmium complexes have been introduced as DNA labels for different applications.4-8 Syntheses of nucleobase conjugates with various metal complexes and techniques of DNA probe labeling have been reported by a number of laboratories.9-12 In analogy to the multicolor optical labeling employed in the DNA sequencing techniques13 as well as in the gene ”chips” (arrays),14 several approaches employing electroactive labels differing in their redox potentials were proposed. For example, Kuhr and co-workers utilized four different ferrocene derivatives as DNA tags in a DNA sequencing technique using capillary electrophoresis coupled to a sinusoidal-voltammetric detector.15 Three ferrocene labels complemented with anthraquinone (representing a tag for the fourth DNA base) were used by Di Giusto et al. who reported on detection of primer extension reactions on self-assembled monolayer-modified gold electrodes.16 Weizman and Tor recently introduced4 “tunable” electroactive DNA tags based on Ru or Os complexes involving negatively charged ligands (L-) such as acetylacetonate or hydroxamate. The [Me(bipy)2L-]2+ complexes (bipy ) 2,2′-bipyridine) exhibited substantially lower oxidation potentials than those of analogous [Me(bipy)3]2+ complexes. The authors proposed utilization of the redox-tunable complexes as potential DNA labels for the “multicolor” electrochemical DNA detection allowing parallel analysis of multiple target DNAs.

Electrochemical techniques designed for detecting DNA hybridization often employ labeling of DNA oligonucleotide (ODN) probes with electroactive moieties (reviewed in refs 1-3). Nucleic acids possess intrinsic electrochemical activity (reviewed in ref 2). However, introduction of a suitable label in one of the DNA strands (usually the “signaling” or “reporter” probe, RP) that form the hybrid duplex results in better distinction between the target DNA (tDNA) and the probe and, thus, improves sensitivity and selectivity of the analysis. Labeling of the RPs with different

(4) Weizman, H.; Tor, Y. J. Am. Chem. Soc. 2002, 124, 1568-1569. (5) Fojta, M.; Havran, L.; Kizek, R.; Billova, S.; Palecek, E. Biosens. Bioelectron 2004, 20, 985-994. (6) Fan, C. H.; Plaxco, K. W.; Heeger, A. J. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 9134-9137. (7) Wang, J.; Polsky, R.; Merkoci, A.; Turner, K. L. Langmuir 2003, 19, 989991. (8) Radi, A. E.; Sanchez, J. L. A.; Baldrich, E.; O’Sullivan, C. K. J. Am. Chem. Soc. 2006, 128, 117-124. (9) Hocek, M.; Stepnicka, P.; Ludvik, J.; Cisarova, I.; Votruba, I.; Reha, D.; Hobza, P. Chem. Eur. J. 2004, 10, 2058-2066. (10) Hurley, D. J.; Tor, Y. J. Am. Chem. Soc. 2002, 124, 3749-3762. (11) Mukumoto, K.; Nojima, T.; Takenaka, S. Tetrahedron 2005, 61, 1170511715. (12) Navarro, A. E.; Spinelli, N.; Moustrou, C.; Chaix, C.; Mandrand, B.; Brisset, H. Nucleic Acids Res. 2004, 32, 5310-5319. (13) Zhang, J. Z.; Fang, Y.; Hou, J. Y.; Ren, H. J.; Jiang, R.; Roos, P.; Dovichi, N. J. Anal. Chem. 1995, 67, 4589-4593. (14) Heller, M. J. Annu. Rev. Biomed. Eng. 2002, 4, 129-153. (15) Brazill, S. A.; Kim, P. H.; Kuhr, W. G. Anal. Chem. 2001, 73, 4882-4890. (16) Di Giusto, D. A.; Wlassoff, W. A.; Giesebrecht, S.; Gooding, J. J.; King, G. C. J. Am. Chem. Soc. 2004, 126, 4120-4121.

* Corresponding author. Tel. +4205 41517197, fax +4205 41211293, e-mail: [email protected]. (1) Drummond, T. G.; Hill, M. G.; Barton, J. K. Nat. Biotechnol. 2003, 21, 11921199. (2) Palecek, E.; Fojta, M. In Bioelectronics; Wilner, I., Katz, E., Eds.; Wiley VCH: Weinheim, 2005, pp 127-192. (3) Wang, J. In Electrochemistry of nucleic acids and proteins. Towards electrochemical sensors for genomics and proteomics; Palecek, E., Scheller, F., Wang, J., Eds.; Elsevier: Amsterdam, 2005; pp 369-382.

1022 Analytical Chemistry, Vol. 79, No. 3, February 1, 2007

10.1021/ac0616299 CCC: $37.00

© 2007 American Chemical Society Published on Web 12/22/2006

Another approach has been introduced by Wang et al. who proposed17 a DNA coding technology based on sulfide nanoparticles of metals differing in their redox potentials (zinc, lead, cadmium). Probably the first electrochemical DNA markers, based on complexes of osmium tetroxide with tertiary amine ligands (Os,L), were introduced18,19 by one of us (E.P.) at the beginning of 1980s. It has been shown19-22 that the Os,L form stable covalent adducts predominantly with thymine residues. Among other DNA bases, cytosine exhibits about 10-times lower reactivity toward Os,L while purine bases are practically unreactive.21,22 Modification of thymine residues in DNA by some of the Os,L (such as a complex of osmium tetroxide with 2,2′-bipyridine; Os,bipy) is sensitive to DNA structure, exhibiting a considerable selectivity for single-stranded DNA.19 Osmium-modified DNA produces well-defined and analytically useful voltammetric signals at carbon,5,22,23 mercury5,18,21,24 or amalgam electrodes.25 The Os,bipy complex has recently been utilized for covalent labeling of oligonucleotide probes5,25 or target DNAs23,26,27 in electrochemical DNA hybridization assays, including techniques of determination of the length of repetitive DNA sequences.5,26,27 Here we show that considerable shifts of peak potentials (at the carbon electrode) of DNA-Os,L can be achieved by replacing bipy with other nitrogenous ligands (L). In this way, the technology of RP labeling with the Os,L can be used for electrochemical “multicolor “ (multipotential) DNA coding. We also for the first time show that this technology can be used for labeling of oligonucleotide probes hybridizing with target DNAs that contain all four DNA bases. EXPERIMENTAL SECTION Material. Synthetic oligonucleotides (ODNs) were purchased from VBC-Genomics (Austria): reporter probes (RPs) (GAA)7(T)20, (RP1); (GGGAAAGAGAGGGAAAGAGAG)7(T)20 (RP2); (CGG)7(T)20 (RP3); GGATGGGCCTCCGGTTCATGAA(T)20 (RP4); target DNAs (tDNAs) (TTC)7(A)20 (tDNA1); CTCTCTTTCCCTCTCTTTCCC(A)20 (tDNA2);CTCTCTTTCCCTCTCTTTCCCGGG(TTC)7(A)20 (tDNA12); (CCG)7(A)20 (tDNA3); CATGAACCGGAGGCCCATCC(A)20 (tDNA4); 5′-biotinylated ODN CATGAACCGGAGGCCCATCC (comp4). Each RP comprised a T20 tail to be covalently modified with the osmium tetroxide complexes, and a segment recognizing complementary sequence in the corresponding tDNA. Each tDNA involved, besides the segment(s) forming a hybrid duplex with a complementary RP, an A20 stretch serving as an adaptor for capture of the tDNAs at the magnetic beads (17) Wang, J.; Liu, G.; Merkoci, A. J. Am. Chem. Soc. 2003, 125, 3214-3215. (18) Palecek, E.; Hung, M. A. Anal. Biochem. 1983, 132, 236-242. (19) Palecek, E. Methods Enzymol. 1992, 212, 139-155. (20) Palecek, E.; Vlk, D.; Vojtiskova, M.; Boublikova, P. J. Biomol. Struct. Dyn. 1995, 13, 537-546. (21) Jelen, F.; Karlovsky, P.; Pecinka, P.; Makaturova, E.; Palecek, E. Gen. Physiol. Biophys. 1991, 10, 461-473. (22) Fojta, M.; Havran, L.; Kizek, R.; Billova, S. Talanta 2002, 56, 867-874. (23) Fojta, M.; Havran, L.; Billova, S.; Kostecka, P.; Masarik, M.; Kizek, R. Electroanalysis 2003, 15, 431-440. (24) Havran, L.; Fojta, M.; Palecek, E. Bioelectrochemistry 2004, 63, 239-243. (25) Yosypchuk, B.; Fojta, M.; Havran, L.; Heyrovsky, M.; Palecek, E. Electroanalysis 2006, 18, 186-194. (26) Fojta, M.; Havran, L.; Vojtiskova, M.; Palecek, E. J. Am. Chem. Soc. 2004, 126, 6532-6533. (27) Fojta, M.; Brazdilova, P.; Cahova, K.; Pecinka, P. Electroanalysis 2006, 18, 141-151.

Figure 1. Formation of covalent adduct of a DNA thymine residue and the Os,bipy complex; ligands that can replace bipy in the complex/ adduct: 1,10-phenanthroline (phen); bathophenanthroline disulfonic acid (bpds); 2,9-dimethyl 1,10-phenanthroline (neocuproine, neoc); N,N,N′,N′-tetramethylethylenediamine (TEMED, tem). Nitrogen atoms taking part in coordination of the osmium atom are highlighted by red boxes.

(see below). Dynabeads Oligo(dT)25 (DBT) were supplied by Dynal A.S. (Norway). Osmium tetroxide was purchased from JMC (UK), 2,2′-bipyridine (bipy), 1,10 phenanthroline (phen), bathophenanthroline disulfonic acid (bpds), 2,9-dimethyl 1,10-phenanthroline (neocuproine, neoc), and N,N ,N ′,N ′-tetramethyl ethylenediamine (TEMED, tem) from Sigma (for structures, see Figure 1). All reagents were of analytical grade. Reporter Probe Labeling with Osmium Tetroxide Complexes. (a) For modification of RP1 or RP2 with Os,bipy, Os,phen, Os,bpds, or Os,neoc, the following procedure was used: the RP (3.8 µM) was incubated with the given Os,L (2 mM) in 100 mM Tris buffer (pH 7.4) at 37 °C for 6 h followed by removal of unreacted Os,L by dialysis against the same buffer using SlideA-Lyzer MINI Dialysis Units (Pierce, Rockford, IL) at 5 °C overnight. (b) RP1 or RP2 labeling with Os,tem was conducted as above, but 3 mM osmium tetroxide mixed with 5 mM TEMED was used. (c) Labeling of RP3 with Os,bipy or Os,neoc was performed under the following conditions: 2 mM Os,L at 20 °C for 2 h, followed by dialysis as in method a. (d) RP4 was first annealed with the biotinylated comp4 ODN [3.8 µM each strand in 0.3 M NaCl, 10 mM Tris-HCl, pH 7.6 (buffer H); heated to 75 °C and slowly cooled down to 20 °C], followed by incubation with 2 mM Os,bipy at 20 °C, 2 h in the same medium. After removal of unreacted Os,bipy by dialysis (against the buffer H), the ODN duplexes were captured at streptavidin-coated magnetic beads, washed, transferred into 10 mM Tris/1 mM EDTA (pH 7.8) (TE buffer), and the labeled RP4(Os,bipy) strand was released by heating at 75 °C for 2 min. Concentrations of Os,L-modified RPs were determined voltammetrically.22 Electrochemical Analysis of Os,L-Labeled Reporter Probes. Electrochemical responses of the RP(Os,L)s were measured by means of “ex-situ” adsorptive transfer stripping (AdTS) squarewave voltammetry (SWV) at a pyrolytic graphite electrode (PGE) using a CHI 440 electrochemical analyzer (CH-Instruments, Austin, TX). Briefly, the modified RPs were adsorbed at the PGE surface from 5-µL drops of solution containing 0.2 M NaCl/10 Analytical Chemistry, Vol. 79, No. 3, February 1, 2007

1023

Figure 2. Scheme of the RP labeling with Os,L. The probes comprise a recognition segment complementary to the tDNA (underlined) and a T20 stretch (red) to be covalently modified with the Os,L. (A) Probes lacking thymine residues within the recognition segment: (i) purine bases (R) do not react with the Os,L. Probes with homopurine recognition segments can thus be easily labeled exclusively within the T20 tail. (ii) Specific labeling of the T20 tail in a probe involving cytosine residues (which are about ten times less reactive toward Os,L, compared to thymines) within its recognition segment can be achieved under controlled (milder) reaction conditions. (B) Probes involving all four bases within the recognition segment: In such probes, the recognition segment must be protected during the modification reaction. Since the thymine-Os,L reaction is single strand-selective when ligands such as bipy, tem or bpds are used, the protection can be achieved through duplex formation with an ODN complementary to the recognition segment (while the T20 tail comprises a single-stranded overhang). After modification, the protecting ODN (end-labeled with biotin) can easily be separated from the RP(Os,L) using magnetic beads coated with streptavidin (DBstv).

mM Tris-HCl (pH 7.4) for 60 s, followed by double rinsing with deionized water and dipping into the voltammetric cell containing blank background electrolyte (0.2 M sodium acetate, pH 5.0). The SWV measurements were performed with the following settings: initial potential -1.0 V, quiescent time 2 s, pulse amplitude 25 mV, frequency 200 Hz, potential step 5 mV, final potential +0.1 V (a procedure previously optimized for measurements of DNA modified with Os,bipy22). DNA Hybridization at Magnetic Beads. The double-surface DNA hybridization assay was performed as follows: 20 µL of tDNA solution (or of mixtures of different tDNAs) in the buffer H was added to the DBT (10 µL of the stock suspension per sample) 1024 Analytical Chemistry, Vol. 79, No. 3, February 1, 2007

previously washed twice by 50 µL of the buffer H. The mixture was shaken in the Thermomixer Comfort (Eppendorf) for 30 min at 20 °C to allow hybridization between the tDNA A20 adaptors and the T25 stretches immobilized on the beads surface. The beads were then washed three times by 50 µL of the buffer H, followed by addition of 20 µL of RP(Os,L) solution (or mixture of different probes) in the buffer H. To allow the tDNA-RP hybridization, the DBT suspension was incubated in the Thermomixer as above. After triplicate washing with 50 µL of the buffer H, the beads were resuspended in 10 µL of TE buffer, the hybrid duplexes were denatured by heating at 85 °C for 2 min, and the released RP(Os,L) was determined by the AdTS SWV measurements (see

Figure 4. Dependence of intensities of the Os,L label signals on the RP1(Os,L) concentration: (circles) peak βbipy yielded by RP1(Os,bipy); (squares) peak βneoc yielded by RP1(Os,neoc); (triangles) peak βtem yielded by RP1(Os,tem); (crosses) unmodified RP1. For other details, see Figure 3B.

Figure 3. (A) Square-wave voltammograms of RP1 modified with various Os,Ls: (1) background electrolyte; (2) RP1 treated with OsO4 alone; (3) RP1(Os,bipy); (2) RP1(Os,phen); (3) RP1(Os,bpds); (4) RP1(Os,neoc); (5) RP1(Os,tem). All curves were measured for RP1(Os,L) concentration of 380 nM. The voltammograms were measured with initial potential -1.25 V (curves 3 and 7) or -1.0 V (others). (B) Sections of baseline-subtracted square-wave voltammograms at PGE of (1) RP1(Os,bipy); (2) RP1(Os,neoc); (3) RP1(Os,tem); initial potential -1.0 V, other conditions as in part A. Table 1. Potentials of Square-Wave Voltammetric Peaks of RP1 Labeled with Various Os,Lsa

peak βL peak RL

Os,bipy

Os,phen

Os,bpds

Os,neoc

Os,tem

-0.080 -0.580

-0.125 -0.545

-0.130 -0.530

-0.185 -0.610

-0.365 -0.975b

a Data obtained for 380 nM RP1(Os,L), accumulation time 60 s, potential of reduction/ initial potential Ei ) -1.0 V. b Same as above but with Ei ) -1.25 V.

above; NaCl was added to each sample to a final concentration of 0.2 M prior to adsorption at the PGE). In all steps, the beads were separated from the supernatant using the magnetic concentrator and resuspended in a new medium by short vortexing.

Figure 5. Scheme of the electrochemical double-surface (DS) DNA hybridization assay. From top to bottom: target DNA (tDNA) is captured at magnetic Dynabeads oligo(dT)25 (DBT) via DNA duplex formation between T25 chains covalently attached to the DBT surface and A20 stretches present in the target ODNs. After magnetic separation and washing of the beads, a RP(Os,L) is hybridized with the tDNA. The DBT-tDNA-RP(Os,L) “sandwich” is formed only in the case of nucleotide sequence complementarity between tDNA and the RP. After another washing step, the hybrids are denatured by heat, the DBT removed, and the released RP(Os,L) determined using AdTS SWV as in Figure 3.

RESULTS AND DISCUSSION 1. Labeling of Reporter Probes Oligo(T) Tails with the Os,L. Os,bipy-labeled reporter probes were recently designed by us5 for hybridization analysis of a repetitive sequence (GAA)n· Analytical Chemistry, Vol. 79, No. 3, February 1, 2007

1025

Figure 6. (A) Dependence of intensities of the RP(Os,L) signals, resulting from the electrochemical DS hybridization assay, on concentration of target DNA: (circles) tDNA1 and RP1(Os.bipy), peak βbipy; (squares) tDNA1 and RP1(Os.neoc), peak βneoc; (triangles) tDNA1 and RP1(Os.bipy), peak βtem; (crosses) tDNA2 and RP1(Os.bipy); RP concentration 380 nM. (B) Comparison of intensities of peak Rbipy resulting from DBT hybridization of Os,bipy-modified RPs with complementary tDNAs (indicated below the graph). The RP(Os,bipy)s were prepared (a) as in Figure 2Ai, (b) as in Figure 2Aii, (c) as in Figure 2B. Concentration of the tDNAs was 150 nM, the RP’s concentration 380 nM; for other details, see Figures 3B and 5.

(TTC)n (related to a neurodegenerative disease Friedreich ataxia28). Since purine bases are practically unreactive toward the Os,bipy,21,22 a RP (GAA)nTx can easily be labeled specifically within the Tx tail (overhang) without modification of the (GAA)n segment recognizing the target DNA. For modification of the two homopurine reporter probes (RP1 and RP2, Figure 2Ai) with Os,bipy, Os,phen, Os,bpds, or Os,neoc, reaction conditions previously optimized5 for probe labeling with Os,bipy were used. For Os,tem which, compared to the former complexes, exhibits weaker reactivity toward the thymine residues (M. Trefulka and E. Palecˇ ek, unpublished), the reaction conditions were empirically adapted to achieve maximum probe labeling (see Experimental Section). In addition to the above RPs, we also used probes containing either cytosines (RP3) or both cytosines and thymines (RP4) within the tDNA-recognizing segments. Labeling of the RP3 with Os,bipy required using milder reaction conditions (Figure 2Aii) (28) Campuzano, V.; Montermini, L.; Molto, M. D.; Pianese, L.; Cossee, M.; Cavalcanti, F.; Monros, E.; Rodius, F.; Duclos, F.; Monticelli, A.; et al. Science 1996, 271, 1423-1427.

1026 Analytical Chemistry, Vol. 79, No. 3, February 1, 2007

to prevent modification of cytosine residues which are approximately 10-times less reactive toward Os,bipy than thymines.21,22 Incubation of the RP3 with 2 mM Os,bipy at 20 °C for 2 h was sufficient to achieve about 90% of modification, compared to RP1(Os,bipy) modified as in Figure 2Ai [calculated from intensity of voltammetric peak Rbipy (Figure 3) measured at PGE for 380 nM RP(Os,bipy)]. When an ODN (CCG)7 (lacking the T20-tail) was treated in the same way, the resulting voltammetric signal was negligible, suggesting that cytosine residues were practically unmodified (not shown). This was confirmed by testing the RP3(Os,bipy) hybridization capacity (see below). RP3 labeled with Os,neoc was prepared in an analogous way. Labeling of the RP4 containing all four bases via incubation of the single-stranded ODN with Os,bipy could not be used because such treatment would result in modification of thymine residues within the recognition segment. In our previous work, DNA modification with Os,bipy was utilized to prevent undesired DNA renaturation;23,26,27 the same treatment can thus be expected to abrogate hybridization between a tDNA and a Os,L-modified RP possessing thymines within the complementary stretch. Nevertheless, it is known that reaction of thymine residues in doublestranded DNA with Os,bipy is sterically hindered.19,22 Similar selectivity for single-stranded DNA is exhibited by some other Os,Ls such as Os,tem or Os,bpds (but not by Os,phen20). This makes it possible to protect thymines within the RP recognition segment during reaction with these Os,Ls via DNA duplex formation with a complementary ODN [while keeping the probe oligo(T) tail single-stranded, Figure 2B]. We annealed the RP4 with a 3′-biotinylated ODN comp4, followed by treatment with Os,bipy as above. For separation of the RP4(Os,bipy) from the comp4 strand, the modified duplex was captured at magnetic beads coated with streptavidin, and the labeled probe was released by thermal denaturation. The degree of RP4 Os,bipy modification attained in this way was comparable to that reached with the single-stranded RP3, and, importantly, the RP4(Os,bipy) retained its ability to hybridize with a complementary tDNA4 (see below). 2. Effect of Ligand Exchange on the RP(Os,L) Electrochemical Behavior. The RP1 was treated with Os,Ls involving different ligands: bipy, phen, bpds, neoc, or tem (Figure 1), and voltammetric responses of the modified ODNs at PGE were measured. An earlier optimized ex-situ voltammetric procedure was used,5,22 involving adsorption of the RP(Os,L) at the electrode surface, electroreduction of the osmium adducts at -1.0 V, and measurement of signals due to anodic reoxidation of the osmium species. After the RP1 treatment with osmium tetroxide alone, no osmium signal was obtained (Figure 3A) in agreement with the fact that thymine glycol but no stable osmium adduct is formed as the final reaction product in the absence of nitrogenous ligands.19 RP1 treated with Os,bipy, Os,phen, Os,bpds, or Os,neoc yielded two well-defined signals in the potential region between -1.0 and +0.1 V (Figure 3; according to ref 22 we denominated these signals as “peak RL” and “peak βL” where L refers to the given ligand) under the given conditions. Potentials of the respective peaks differed depending on the ligand type (among the above Os,Ls, peak RL potentials varied between -0.61 and -0.53 V and peak βL potentials between -0.08 and -0.185 V, Table 1). RP1(Os,tem) produced only one signal within the same region (peak βtem at -0.365 V). The more negative peak

Figure 7. Parallel detection and distinction among three target nucleotide sequences using Os,L-labeled reporter probes. Target DNAs were captured at the DBT: (i), tDNA1; (ii), tDNA2; (iii), tDNA12; (iv) tDNA3. Then, the beads were incubated with a mixture of Os,L-labeled RPs: RP1(Os,bipy), RP2(Os,tem) and RP3(Os,neoc), followed by voltammetric detection of captured RP(Os,L). Target DNA concentrations were 150 nM, RP concentration 230 nM of each.

Rtem appeared at -0.975 V only when the initial potential was shifted to more negative values. With RP1(Os,bipy), shifting the initial potential to Ei)-1.25 V did not result in substantial changes in shape and intensity of its peaks, and no additional signal more negative than the peak Rbipy appeared (Figure 3A). The best peak separation was achieved between signals of RP1(Os,tem) and those of RP1(Os,bipy), RP1(Os,phen), or RP1(Os,bpds). On the other hand, the latter three labels could not be well distinguished due to rather small differences in their redox potentials (Table 1). We thus chose three Os,Ls offering reasonable separation of each of their peaks βL: Os,tem, Os,neoc, and Os,bipy (Figure 3B). Signals of the RP1(Os,bipy) and RP(Os,tem) could be measured simultaneously as they only overlapped slightly. Simultaneous detection of other combinations of the Os,Llabeled probes was limited due to more significant signal overlaps, but the peak separations still allowed unambiguous identification of the given label. Using RP1 labeled with Os,bipy, Os,neoc, or Os,tem, we measured the RP1(Os,L) voltammetric responses as a function of its concentration. Intensities of the RP1(Os,L) signals increased linearly with the RP1 concentration within the region of 7.6 to 380 nM (in Figure 4 shown for peak βbipy, βneoc and βtem yielded by RP1 labeled with Os,bipy, Os,neoc, or Os,tem, respectively). Under the given conditions, the lowest detectable concentration for RP1(Os,bipy) was 3.8 nM while for RP1(Os,neoc) or RP1(Os,tem) about 6 nM. Considering the sample volume (5 µL), it was

possible to detect 19 fmol of RP(Os,bipy) or about 30 fmol of RP1(Os,neoc) or Os(tem), respectively, after a 60-s accumulation time. Unmodified RP1 (or any other unmodified ODN) yielded no signal within the potential region of -1.25 to +0.1 V for any RP1 concentration. 3. Double-Surface Electrochemical DNA Hybridization Assay with the Os,L-Labeled Probes. For the DNA hybridization experiments, we used the double-surface (DS) technique involving target DNA capture at the magnetic Dynabeads oligo(dT)25 (DBT) (Figure 5). We5,23,26,29-31 and others7,17,32 (reviewed in refs 2, 3, 33) have recently applied the “biomagnetic” technology in various assays of DNA hybridization or DNA protein interactions, including those involving the Os,bipy DNA marker.5,23,26,29 For this work, we designed model target ODNs with A20 adaptors allowing their immobilization at the DBT: tDNA1 (derived from trinucleotide repetitive sequence related to a hereditary disease Friedreich ataxia26,28), tDNA2 (containing a random homopyrimidine stretch), tDNA3 (derived from a triplet repeat related to the fragile X chromosome syndrome34), tDNA4 (section of a coding (29) Palecek, E.; Kizek, R.; Havran, L.; Billova, S.; Fojta, M. Anal. Chim. Acta 2002, 469, 73-83. (30) Palecek, E.; Fojta, M.; Jelen, F. Bioelectrochemistry 2002, 56, 85-90. (31) Palecek, E.; Masarik, M.; R., K.; Kuhlmeier, D.; Hassmann, J.; Schu ¨ lein, J. Anal. Chem. 2004, 76, 5930-5936. (32) Wang, J.; Xu, D. K.; Erdem, A.; Polsky, R.; Salazar, M. A. Talanta 2002, 56, 931-938.

Analytical Chemistry, Vol. 79, No. 3, February 1, 2007

1027

Figure 8. DS hybridization analysis of tDNA1 and tDNA2 mixed at various ratios using Os,L-labeled RPs. Mixture of tDNA1 and tDNA2 was captured at the DBT, followed by hybridization with a mixture of RP1(Os,bipy) and RP2(Os,tem). (A) The section of voltammograms resulting from the hybridization experiments. Target DNA concentrations: (red curve) 15 nM tDNA1 and 60 nM tDNA2; (blue curve) 37.5 nM of each; (black curve) 60 nM tDNA1 and 15 nM tDNA2 (total tDNA concentration was always 75 nM); RP concentrations: 150 nM of each. (B) Dependence of intensities of specific signals of two Os,L RP labels on the tDNA2/ tDNA1 ratio: (red circles) peak βtem due to RP2(Os,tem); (black triangles) peak βbipy due to RP1(Os,bipy).

sequence of tumor suppressor gene p5335), and tDNA12 (bearing the specific sequences of tDNA1 and tDNA2 in a single strand). The hybridization experiment is depicted on Figure 5. Briefly, a tDNA (or mixture of tDNAs) was incubated with the DBT to allow hybridization between the T25 chains bound to the bead surface and the A20 adaptors in the tDNAs. After washing, the beads were incubated with a Os,L-labeled RP (or mixture of RPs) followed by another washing and release of hybridized species by thermally induced DNA denaturation. The RP Os,L labels were determined voltammetrically as described above. Testing Hybridization Capacity of the Os,L-Labeled RPs. It has been shown previously5 that the presence of an Os,bipy-modified oligo(T) tail consisting of up to 30 thymines did not significantly (33) Wang, J. In Electrochemistry of nucleic acids and proteins. Towards electrochemical sensors for genomics and proteomics; Palecek, E.; Scheller, F., Wang, J.; Eds.; Elsevier: Amsterdam, 2005; pp 175-190. (34) Paulson, H. L.; Fischbeck, K. H. Annu. Rev. Neurosci. 1996, 19, 79-107. (35) Selivanova, G. Curr. Cancer Drug Targets 2004, 4, 385-402.

1028 Analytical Chemistry, Vol. 79, No. 3, February 1, 2007

affect hybridization between the the RP and tDNA trinucleotide repeat (GAA)7·(TTC)n. Intensity of voltammetric signals resulting from the DS hybridization assay increased linearly with the oligo(T) tail length (because of accumulation of the Os,bipy moieties) as well as with the concentration of the tDNA (while the RP was applied in an excess), suggesting that Os,bipy labeling of the oligo(T) tails did not affect the RP ability to form hybrid duplexes with the tDNA.5 Here we tested whether or not modification of the RPs with different Os,Ls affects their hybridization capacity. We modified RP1 with Os,bipy, Os,neoc, or Os,tem and used these probes in the DS DNA hybridization assay. Voltammetric responses resulting from these experiments gave evidence that RP1 modified with any of the Os,Ls efficiently hybridized with the DBTcaptured tDNA1. Signals obtained after the DS procedure were specific for the label used for the RP1 modification and exhibited linear dependence of their intensities on tDNA1 concentration (in the solution from which the tDNA1 was captured at the DBT) within the range of 15 to 150 nM (while the RPs were applied at a constant concentration of 230 nM, Figure 6A). The target DNAs were easily detectable using the Os,L-labeled RPs down to 4-7 nM (corresponding to 80-140 fmol of tDNA in a 20-µL sample used for the DS hybridization assay). When tDNA2 (noncomplementary to the RP1) was used instead of tDNA1, no distinct signal was detected (Figure 6A). Analogous results were obtained also with RP1 labeled with Os,phen and Os,bpds (data not shown). We further tested effects of Os,bipy treatment on hybridization capacity of different RPs. First, RP1, RP2, RP3, and RP4 were modified in their single-stranded forms as in Figure 2Ai (2 mM Os,bipy, 37 °C, 6 h). Under these conditions, signals resulting from hybridization of the Os,bipy-labeled “homopurinic” RPs (RP1 or RP2) with the complementary tDNAs (tDNA1 or tDNA2, respectively) exhibited similar intensities (Figure 6B). When RP3 was treated with Os,bipy in the same way, the intensity of signal resulting from its hybridization with tDNA3 was only about 50%, compared to signal resulting from RP1(Os,bipy)‚tDNA1 hybridization (Figure 6B). On the other hand, measurement of the RP3(Os,bipy) voltammetric response without the DBT hybridization procedure revealed at least the same modification degree as that attained with RP1 or RP2 (not shown). A (CCG)7 ODN (lacking the T20 tail) treated with Os,bipy under the same conditions yielded distinct signals of the DNA-Os,bipy adducts; compared to the signal yielded by the Os,bipy-modified (TTG)7 ODN, the peak Rbipy due to (CCG)7(Os,bipy) was about 25%, suggesting that 3 to 4 cytosine residues of 14 in the (CCG)7 repeat might be modified. Thus, it was the cytosine-(Os,bipy) adducts within the RP3 recognition segment which caused the observed reduction of the RP3‚tDNA3 hybridization signal. When the RP3 was labeled with Os,bipy using milder conditions (as in Figure 2Aii) under which no modification of the (CCG)7 ODN was detected (see above), the intensity of the hybridization signals was close to 90% of the signals obtained with RP1(Os,bipy)‚tDNA1 or RP2(Os,bipy)‚ tDNA2. When RP4 (involving all four bases in its recognition segment) was treated with Os,bipy in its single-stranded form (as in Figure 2Ai), only a negligible signal resulted from the RP4(Os,bipy)‚ tDNA4 hybridization (Figure 6B). On the other hand, RP4(Os,bipy) labeled in the duplex with the biotinylated comp4 protection ODN (Figure 2B) yielded a similar hybridization signal as that

obtained with the RP3(Os,bipy)‚tDNA3. All Os,bipy-labeled RPs prepared using the appropriate procedures yielded responses linearly increasing with the complementary tDNA concentration in the same range as observed with the Os,L-labeled RP1 (Figure 6A). Similar behavior was exhibited by other Os,L-modified RPs tested (after hybridization with complementary tDNAs): RP2(Os,tem), RP2(Os,neoc), and RP3(Os,neoc). Moreover, any noncomplementary RP‚tDNA pair produced no significant response (data not shown). These results suggest that (1) any of the Os,Llabeled RPs could efficiently recognize the complementary tDNA and form a stable hybrid, and (2) the RP(Os,L)-tDNA recognition was highly selective, without false positive signals. 4. Parallel Detection of Target DNAs Using Different Os,L Probe Labels. We first captured at the DBT one of the following tDNAs: tDNA1, tDNA2, tDNA12, or tDNA3, followed by hybridization with a 1:1:1 mixture of RP1(Os,bipy), RP2(Os,tem), and RP3(Os,neoc). The scheme of this experiment and the voltammetric responses obtained are shown in Figure 7. Electrochemical responses resulting from this procedure corresponded to the Os,L used for labeling of probes recognizing the given tDNA. With target ODNs bearing nucleotide sequences complementary to only one of the RPs (Figure 7i, ii, and iv), distinct signals specific for the particular Os,L labels were obtained. For tDNA1 complementary to the RP1, two signals at the potentials matching those of peaks Rbipy and βbipy were detected (Figure 7i). When tDNA2 complementary to RP2 was used, a single signal corresponding to the peak βtem was observed (Figure 7ii) while for tDNA3 hybridizing with the RP3, a voltammogram corresponding to the Os,neoc label was obtained (Figure 7iv). Finally, when the tDNA12 involving stretches complementary to RP1 and RP2 was captured at the DBT, the resulting voltammogram displayed signals of Os,bipy and Os,tem labels, suggesting that both RPs hybridized simultaneously with this tDNA (Figure 7iii). Intensities of the signals measured after hybridization with the 65-mer tDNA12 (Figure 7iii) were about 16% lower, compared to the respective peaks resulting from the same experiment with the 41-mer tDNA1 (Figure 7i) or tDNA2 (Figure 7ii). This effect might be caused by less efficient capture of the longer tDNA at the beads and/or by mutual interference of the two RPs simultaneously hybridizing with the tDNA2 [albeit the latter effect was probably less significant since our previous study5 with three 97-mer ODNs involving the repeated (TTC)n sequence capable of accommodation of one, two, or three molecules of the RP1(Os,bipy) revealed a linear relation between the number the tDNA repeat units and intensity of the measured signal]. Further, mixtures of tDNA1 and tDNA2 were prepared and captured at the DBT, followed by hybridization with a 1:1 mixture of RP1(Os,bipy) and RP2(Os,tem). Voltammetric responses resulting from this experiment showed well-resolved signals of both Os,L labels for molar ratios tDNA2/tDNA1 ranging from 0.25 to 4 (total tDNA concentration was 75 nM), and relative intensities

of these peaks reflected changes in the tDNA1/tDNA2 ratio (Figure 8). The RP1(Os,bipy) yielded well developed peaks even when the tDNA1 was mixed with tDNA2 in a ratio 1:9 (not shown); on the other hand, when the tDNA ratio was inversed, the peak βtem due to RP2(Os,tem) was substantially overlapped by peak Rbipy yielded by RP1(Os,bipy) hybridizing with the excess of tDNA1.

(36) Isono, K.; Niwa, Y.; Satoh, K.; Kobayashi, H. Plant Physiol. 1997, 114, 623630. (37) Fan, Z. H.; Mangru, S.; Granzow, R.; Heaney, P.; Ho, W.; Dong, Q.; Kumar, R. Anal. Chem. 1999, 71, 4851-4859.

Received for review August 31, 2006. Accepted November 13, 2006.

CONCLUSIONS In this paper we show that covalent RP labeling with the osmium tetroxide complexes can be used for “multicolor” electrochemical DNA coding and parallel hybridization analysis of multiple tDNAs. Changes in redox potentials of the Os,L labels due to ligand (L) exchange are sufficient for their reliable distinction and, when suitable Os,Ls are chosen, even for simultaneous detection. One of the major advantages of this DNA labeling approach is its feasibility in any biochemical or molecular biology laboratory. Modification of DNA with Os,L takes place in aqueous media at physiological conditions using millimolar concentrations of the reagents5,19,22 and does not require any specialized organic chemistry equipment and/or personnel which is usually necessary for synthesis of metal complex-conjugated nucleobases, nucleotide triphosphates, or ODNs.4,9,10 Labeling of any ODN composed of standard bases with electroactive markers can be done instantly when needed. The Os,L-labeled ODNs are stable and can be stored for several months in a freezer. Multiple Os,L labels on the oligo(T) stretch inherently offer a better sensitivity of the probe detection, compared to techniques using a single label per probe molecule.5 We also demonstrate that utilization of this technology is not restricted to homopurinehomopyrimidine sequences, but it can be applied for labeling of any (mixed-sequence) RP with the oligo(T) tail when appropriate modification procedure is used. We present here a study based on model synthetic oligonucleotides; nevertheless, we have successfully applied an analogous technique to analyze PCRamplified natural genomic DNA sequences such as fragments of human tumor suppressor gene p5335 or a plant gene rbcL36 [results of the RP(Os,L) hybridization with the PCR amplicons will be published elsewhere]. The concept based on the magnetic beads can be multiplexed, and it is compatible with nanofluidic technology37 which makes it promising for further development in bioanalytical techniques and its spread in medical practice. ACKNOWLEDGMENT This work was supported by GA ASCR (IAA4004402), GACR (203/05/0043, 203/ 04/1325), Ministry of Industry and Trade of the CR (1H-PK/42), Ministry of Education, Youth and Sports of the CR (LC06035), and by a Institutional Research Plan No. AVOZ50040507. The authors thank Mrs. Iva Salajkova´ for technical assistance.

AC0616299

Analytical Chemistry, Vol. 79, No. 3, February 1, 2007

1029