Biomolecule detection with peroxidase-mimicking DNAzymes; expanding detection modality with fluorogenic compounds

Abstract

The identification of a suitable non-fluorescent molecule that can be oxidized by G-quadruplexes into a fluorescent product will be important for several fields such as bioanalyte sensing and cancer therapeutic discovery. Herein, we demonstrate that 2′,7′-dichlorodihydrofluorescein diacetate is a superior reducing substrate for the fluorometric detection of bioanalytes using peroxidase-mimicking G-quadruplexDNAzymes.

---

G-quadruplex DNAzymes ¹ have emerged as an important, label-free and cheap alternative to the use of horseradish peroxidases (HRP) for the detection of bioanalytes, such as DNA/RNA,²proteins,³nucleotides⁴ and even toxic transition metals.⁵ In contrast to protein labels, the use of nucleic acidcatalysts for the detection of bioanalytes eliminate the need for strict refrigerated storage as most oligonucleotidecatalysts can be stored as dry solids and reconstituted into active catalysts in water when the need arises.

For the detection of bioanalytes using G-quadruplexDNAzymes, 2,2′-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) diammonium salt (ABTS²⁻) have predominantly been used as chromogenic substrates (Fig. 1).² The oxidized product of ABTS is a radical cation, which can simply be detected via its absorbance at 415 nm using a UV-visible spectrophotometer. The detection of bioanalytes via absorbance methods has several advantages; such as ease of operation and the lower cost of UV-visible spectrophotometers compared with fluorometers or photon counting instruments. On the downside, detection methods that rely on absorbance tend to have lower dynamic range and sensitivity (absorbance on most common UV-visible spectrophotometers is rarely linear after 3.5 units). Additionally, the presence of other biomolecules such as cytochromes, which also absorb light in the blue region, can complicate the use of ABTS in assays that use whole cells or crude cell lysates.

Fig. 1 Oxidation of chromogenic compound ABTS, by peroxidase-mimicking G-quadruplexcatalyst G1;⁶ 5′-GGGTAGGGCGGGTTGGGT-3′.

The oxidations of several fluorogenic substrates by HRP into fluorescent molecules have been well documented.⁷ On the other hand, the use of fluorogenic compounds for assaying peroxidase-mimicking DNAzymes has not been extensively investigated.⁸ In light of the increasing number of reports that use G-quadruplexDNAzymes as biolabels,² we initiated a program that aimed to evaluate several fluorogenic substrates that could potentially replace ABTS in biomolecule detection protocols. It was anticipated that insights gained from this study would facilitate a more sensitive detection of G-rich nucleic acids or analytes such as H₂O_2.⁹ We therefore embarked on a study aimed at identifying non-fluorescent substrates (Fig. 2) that could be oxidized by DNAzymes into fluorescent products.

Fig. 2 Structures of molecules used in study.

The first group of compounds that were studied were phenolic compounds. It has been shown that iron-containing oxidative enzymes such as the DNAzyme PS2.M⁸ and HRP can oxidize phenols to dimeric phenols15 (see Scheme 1).¹⁰ The oxidized products 15 are usually more intensely fluorescent than the starting phenols. For the sensitive detection of peroxidase mimicking DNAzymes or oxidizing agents such as hydrogen peroxide, using fluorogenic compounds, it is desirable that the background reaction (i.e. reaction in the absence of DNAzymes or hydrogen peroxide) is relatively low, compared to the reaction rate in the presence of the analyte of interest. In this regard, fluorogenic compounds that are susceptible to oxidations into fluorescent compounds in the presence of oxygen and light (photo-oxidations)¹¹ are not ideal. Preliminary studies indicated that most phenols were either stable to hydrogen peroxide or a hydrogen peroxide/hemin mixture. On the other hand, the addition of phenolic compounds 1–4 (0.7 mM) to G1/hemin complex in Tris-HCl buffer in the presence of either K⁺ or NH₄⁺ and hydrogen peroxide resulted in the oxidation of the phenols into brightly fluorescent compounds. The evolution of the fluorescent oxidized products was monitored at 410 nm using an excitation wavelength of 320 nm (see Fig. 3).

Scheme 1 Iron(IV)-oxo complex oxidation of phenols into fluorescent diphenols 15.

Fig. 3 Oxidation of phenolic compounds 1–4 in Tris-buffer containing different monovalent cations, K⁺ and NH₄⁺. The rate of G1-catalyzed oxidations of phenolic compounds 1–4 are higher in buffers containing excess NH₄⁺ than those containing excess K⁺. (a) 50 mM Tris-HCl, 50 mM KClbuffer (pH 8.5) and (b) 50 mM Tris-HCl, 150 mM NH₄OAcbuffer (pH 8.5).

For the set of phenolic compounds tested in this study (1–4), tyramine2 was the best phenolic fluorogenic substrate for peroxidase mimicking DNAzymes; the signal-to-noise ratio (i.e. the ratio of fluorescence intensity in the presence of G1DNAzyme and in the absence of G1 after 2 min) in buffers containing excess ammonium cation for tyramine2 was 496 whereas the signal-to-noise ratios for the other phenols1, 3 and 4 were 270, 237 and 129, respectively. We have recently disclosed that although potassium has a higher affinity for G-quadruplexes, peroxidase-mimicking DNAzymes that are formed in the presence of ammonium cation are superior peroxidasecatalysts than those that are reconstituted in the presence of excess potassium or sodium cations.^2b Sen et al. had earlier observed that HEPES–NH₄OHbuffer containing 20 mM KCl was a superior peroxidation catalyst than buffers that lacked ammonium.^1a We have however demonstrated that peroxidation catalysis by the DNAzyme G1 in buffers that contain only ammonium cation are superior to those that contain a mixture of ammonium and potassium cations (the condition that was used by Sen^1a) or only potassium cation.^2b In our previous study, ABTS²⁻ was used as the reducing substrate and we demonstrated that the enhanced peroxidase activity of G-quadruplexes in only ammonium-containing buffer was due to the selective formation of a parallel conformer of G-quadruplex in the presence of hemin i.e.G-quadruplexes with parallel topology are more active peroxidase enzymes than those with antiparallel topology.^2b In our previous study, we however used only ABTS as the reducing agent. Therefore, we wondered if the ammonium cation enhancing effect was a peculiarity for the ABTS substrate or if indeed it was a general effect. In this study, we also found that when phenolic substrates 1–4 are used as reducing agents, DNAzymes that are templated with ammonium cation are better catalysts than those templated by potassium (see Fig. 3 and 4). We therefore conclude, based on results from our earlier study^2b and this current study, that the G1-catalyzed peroxidation rate enhancement observed in buffers containing excess ammonium cation results from interactions of the ammonium cation with the DNAzyme and not with the reducing substrate.

Fig. 4 Comparison of the peroxidation of chromogenic substrate ABTS5 and fluorogenic substrate tyramine2 by peroxidase-mimicking DNAzyme G1 in buffers containing excess ammonium or potassium cation. Both substrates are oxidized faster in buffers that contain excess ammonium cation.

We proceeded to investigate fluorone dyes. Dihydrofluoresceins and dihydrorhodamines are preferred peroxidase substrates over phenols because these dyes have longer excitation and emission wavelengths. Pleasingly compounds 6, 7, 9 and 10 were also found to be good substrates for G-quadruplexperoxidase-mimicking DNAzymes. In our hands, 2′,7′-dichlorodihydrofluorescein diacetate6 emerged as the optimum substrate (amongst the substrates tested; see Fig. 2) for G1-based detection, vide infra.

In order to increase the sensitivity of fluorogenic-based peroxidase-mimicking G-quadruplex sensing, we investigated the origin of the background noise (i.e. fluorescence signal in the absence of G1). We found that the background noise originated from uncomplexed hemin -mediated oxidation of the fluorone compounds 6, 7, 9 and 10. (see Fig. 5).

Fig. 5 The background fluorescence signal from a mixture containing hemin and hydrogen peroxide but not DNA (line b) is higher than that containing hydrogen peroxide and DNA but not hemin (line c) or hemin and DNA but not hydrogen peroxide (line f).

Reactive oxygen species such as hydrogen peroxide have been implicated in several biological processes.¹² Therefore reagents or assays that can detect hydrogen peroxide in biological samples have found utility in biological studies.¹³ In this regard, the Amplex Red–HRP system has emerged as a reagent of choice for the fluorometric detection of hydrogen peroxide in biological samples.¹⁴

We investigated whether a fluorophore–G1 system (such as the 6–G1 system; see Fig. 6) could be used as an alternative to the Amplex Red–HRP system for the detection of hydrogen peroxide. Pleasingly, the 2′,7′-dichlorofluorescein diacetate 6–G1 system can detect low nanomolar concentrations of hydrogen peroxide (Fig. 7); comparable to or even superior to the limit of detection that is achieved with the more expensive Amplex Red–HRP system (see ESI, Fig. S1† ).

Fig. 6 Both ABTS5 and 2′,7′-dichlorodihydrofluorescein diacetate6 can detect G-quadruplex G1 in a concentration dependent manner. [G1]; (a) 1 μM, (b) 500 nM, (c) 50 nM, (d) 10 nM, (e) 500 pM, (f) 0 pM.

Fig. 7 Nanomolar concentrations of H₂O₂ (1 nM) can be detected by fluorogenic compound 6 whereas the limit of detection of the chromogenic substrate ABTS5 is greater than 1 μM. [H₂O₂]; (a) 50 μM, (b) 10 μM, (c) 1 μM, (d) 500 nM, (e) 20 nM, (f) 1 nM, (g) 0 nM.

In conclusion, we have found that most of the fluorogenic substrates that can be converted by horseradish peroxidases are also suitable for peroxidase-mimicking DNAzymes. Although the signal-to-noise ratio of tyramine was the best for the fluorogenic compounds investigated in this study, 2′,7′-dichlorodihydrofluorescein diacetate6 was deemed as the fluorogenic substrate of choice (at least for the set tested¹⁵) for peroxidase-mimicking G-quadruplex-based assays because of its acceptable signal-to-noise ratio and also it is a brighter fluorophore than tyramine. The use of fluorogenic substrates with G-quadruplex labels expands the detection modalities when peroxidase-mimicking DNAzymes are used as biolabels. Therefore for certain scenarios whereby the use of the traditional chromogenic substrate ABTS might not be appropriate, we recommend 2′,7′-dichlorodihydrofluorescein diacetate6 as a superior replacement. We have also confirmed our earlier finding, using fluorogenic substrates, that DNAzymes templated by ammonium cations are better peroxidase enzymes than those templated by potassium cations. The findings in this communication should find utility in the emerging bioanalyte detection technologies that utilize G-quadruplexDNAzymes as a label-free detection platform and also for fluorometric assays to discover anticancer drugs that inhibit telomere or genomic G-quadruplexes.

We thank the University of Maryland, UM graduate research board and NSF for funding our projects.

Notes and references

1. (a) P. Travascio, Y. Li and D. Sen, Chem. Biol., 1998, 5, 505; (b) P. Travascio, P. K. Witting, A. G. Mauk and D. Sen, J. Am. Chem. Soc., 2001, 123, 1337.
2. (a) Z. Cheglakov, Y. Weizmann, M. K. Beissenhirtz and I. Willner, Chem. Commun., 2006, 3205; (b) S. Nakayama and H. O. Sintim, J. Am. Chem. Soc., 2009, 131, 10320; (c) D. M. Kolpashchikov, J. Am. Chem. Soc., 2008, 130, 2934.
3. (a) T. Li, L. Shi, E. Wang and S. Dong, Chem.–Eur. J., 2009, 15, 1036; (b) T. Li, E. Wang and S. Dong, Chem. Commun., 2008, 3654.
4. D. Li, B. Shlyahovsky, J. Elbaz and I. Willner, J. Am. Chem. Soc., 2007, 129, 5804.
5. (a) T. Li, B. Li, E. Wang and S. Dong, Chem. Commun., 2009, 3551; (b) J. Elbaz, B. Shlyahovsky and I. Willner, Chem. Commun., 2008, 1569.
6. The G1 sequence, designed by Willner et al. (see: Y. Xiao, V. Pavlov, T. Niazov, A. Dishon, M. Kotler and I. Willner, J. Am. Chem. Soc., 2004, 126, 7430 ) is a modification of PS2.M (ref. 1).
7. (a) J. E. Whitaker, P. L. Moore, R. P. Haugland and R. P. Haugland, Biochem. Biophys. Res. Commun., 1991, 175, 387; (b) H. H. Gorris and D. R. Walt, J. Am. Chem. Soc., 2009, 131, 6277.
8. One report has documented that phenolic compounds are good fluorogenic substrates for the DNAzyme PS2.M. See: A. M. Rojas, P. A. Gonzalez, E. Antipov and A. M. Klibanov, Biotechnol. Lett., 2007, 29, 227.
9. (a) B. Sljukic, C. E. Banks and R. G. Compton, Nano Lett., 2006, 6, 1556; (b) M. Shao, Y. Shan, N. Wong and S. Lee, Adv. Funct. Mater., 2005, 15, 1478; (c) S. M. De Toledo, M. Haun, E. J. H. Bechara and N. Duran, Anal. Biochem., 1980, 105, 36.
10. A. J. Gross and I. W. Sizer, J. Biol. Chem., 1959, 234, 1611.
11. T. Christ, F. Kulzer, P. Bordat and T. Basche, Angew. Chem., Int. Ed., 2001, 40, 4192.
12. (a) S. Bhattacharjee, Curr. Sci., 2005, 89, 1113; (b) M. Lopez-Lazaro, Cancer Lett., 2007, 252, 1; (c) T. S. Gechev and J. Hille, J. Cell Biol., 2005, 168, 17.
13. (a) S. Bortolami and L. Cavallini, Free Radical Res., 2009, 43, 446; (b) P. C. Panus, R. Radi, P. H. Chumley, R. H. Lillard and B. A. Freeman, Free Radical Biol. Med., 1993, 14, 217.
14. V. Towne, M. Will, B. Oswald and Q. Zhao, Anal. Biochem., 2004, 334, 290.
15. Dihydroethidium 8 was the worst reducing substrate with a S/N of a meager 1.6. Also, for this substrate, the optimum pH was 5.3 and this might explain the low S/N.

---

Footnote

† Electronic supplementary information (ESI) available: Detailed descriptions of experimental procedures. See DOI: 10.1039/b916228c

---