Selective recognition of G-quadruplexes by a dimeric carbocyanine dye

Abstract

A novel dimeric carbocyanine dye is found to recognise G-quadruplex structures selectively compared to mixed sequence or double-stranded DNA molecules. The dye exhibits a preference for G3-bearing parallel quadruplexes and results in a nearly 1400-fold enhancement compared to the absence of DNA and a nearly 30-fold fluorescence enhancement compared to duplex and single stranded DNA molecules. Study of the dye binding with various quadruplex-forming sequences points towards an end-stacking mechanism of interaction. The number of G-quartets and spacer sequences along with the identity of monovalent ions strongly influence the interaction of quadruplexes with the dye.

---

Introduction

The interaction of nucleic acids with cyanine dyes has been the focus of intense research based on myriad applications.^1–5 The ability of such dyes to facilitate staining, imaging and quantization of nucleic acids has played a pivotal role in the revolutionary advances in recombinant molecular biology and DNA technology. The optical properties of cyanine dyes have been leveraged for high-sensitivity detection associated with quantitative polymerase chain reaction (PCR) and flow cytometry.^6,7 The interaction of cyanine dyes with nucleic acids has also been of interest from the perspective of supramolecular multichromophore assembly.⁸ For example, the programmable character of DNA self-assembly and intercalation by cyanine dyes has enabled the creation of fluorescent nanotags.⁹ The structural features of duplex DNA facilitate groove binding and cooperative self-assembly of carbocyanine dyes.^10–12 The sequence traits of the DNA molecules have been used to control the architecture of such assemblies while the length of DNA template has been used to control the dimensions of the aggregate.¹⁰ The intriguing interactions between dicarbocyanine dyes and nucleic acids have even formed the basis of using dye assemblies to control nucleic acid conformations.¹³ The fluorescence spectral properties of cyanine dyes are widely exploited to reveal the presence of nucleic acids.^3,14,15 The most common model of fluorescence-based assaying of dye–nucleic acid interactions relies on quenched or weak fluorescence of free dyes in solution and multifold enhancement in presence of nucleic acids. The quenched state of free cyanine dyes in solution is often related to vibrational losses owing to the mechanical properties of the dyes. The dyes are understood to assume more rigid conformations upon binding nucleic acids thereby leading to enhanced fluorescence.^16,17

The factors underlying effective cyanine dye–duplex DNA interaction have been exploited to develop dyes that are capable of recognizing non-canonical structures of DNA such as G-quadruplexes.^18,19 The reported prevalence of G-rich sequences in the human genome as well as their implication in gene regulation and antitumor activities has made them valuable targets for recognition by small molecules.^20–22 The distinctive groove morphology and π-stacking surface of G-quartets provides interesting opportunities for small molecule interactions.²³ While large numbers of quadruplex-binding ligands have been identified,^24,25 their inadequate selectivity in comparison to duplexes continues to pose a challenge for their effective adoption.^26,27 Conversely, dyes that bind duplex DNA with high affinity need not necessarily bind quadruplex targets with similar affinities.²⁸ A quadruplex-specific dye could serve as a ligand in both a therapeutic as well as a diagnostic context. The latter would rely on a distinctive spectroscopic signature such as substantial fluorescence enhancement of the dye in presence of target quadruplex.^29,30 While, the majority of quadruplex-selective cyanine dyes rely on end-stacking on the terminal quartets, dyes have also been designed that interact with the loops and grooves of the quadruplex target.³¹ Recently, supramolecular aggregates of have also been employed for selective recognition of quadruplexes.^31,32

In this work we describe the use of a dimeric carbocyanine dye 1 (Fig. 1) for selective recognition of G-quadruplexes. The fluorescence of free 1 in solution is quenched based on its propensity to form H-dimers and H-aggregates. The weak emission of free 1 in solution is similar to that of monomeric cyanine dyes in presence of duplex DNA.¹⁰ The dimeric cyanine dye appears to selectively interact with G3 quadruplexes as revealed by a nearly 1400-fold enhancement from absence of DNA and nearly 30-fold enhancement compared to duplex and single stranded DNA molecules. The fluorescence turn-on observed in presence of specific G-rich sequences, is attributed to the deaggregation of the dyes due to selective interaction with quadruplex secondary structures.

Fig. 1 Dimeric carbocyanine dye 1.

Results and discussion

We began our investigation by assessing the interaction of 1 with G-rich oligonucleotides, mixed sequence single stranded and duplex DNA. Dye 1 has the propensity to self-assemble into H-dimers and H-aggregates in the absence of nucleic acids. While such aggregates are fluorescence-quenched,^33,34 interaction with target molecules that disrupt the aggregates is expected to result in a fluorescence turn-on. Further, the extent of dye deaggregation and concomitant fluorescence enhancement would reflect the affinity of the dye for the agent responsible for deaggregation.

Fig. 2 reveals a very large fluorescence enhancement of 1 in presence of the G-rich oligonucleotides. The DNA sequences used in this work are listed in Table 1. In contrast to the quenched fluorescence of 1 alone, progressively higher fluorescence is observed in presence of single-stranded, duplex DNA and various G-rich oligonucleotides. Notably, the enhancement in fluorescence of 1 in presence of G3T-4 is greater by at least an order of magnitude compared to that in presence of single stranded and duplex DNA. The change in aggregation behaviour of 1 upon interaction with G3T-4 is evident from the corresponding UV-visible absorption spectra (see Fig. 3). The absorption maximum of 1 at 430 nm is attributed to H-aggregate formation by the dye.^35,36 These aggregates are fluorescence quenched. The interaction of 1 with G3T-4 and other DNA molecules results in disruption of the H-aggregates, and emergence of H-dimers and monomers that absorb at 530 nm and 550 nm, respectively.³⁷ While H-dimers are fluorescence quenched, the unfolded monomeric conformation of the dye is highly fluorescent.^33,37 G3T-4 produces the most pronounced disruption of H-aggregates of 1 in contrast to other nucleic acid molecules studied. This deaggregation induced mechanism of sensing holds promise especially when the fluorescence enhancement originates due to specific targets. Notably, the monomeric cyanine dye counterpart of 1 is completely unresponsive to the oligonucleotides used (data not shown).

Fig. 2 Comparative fluorescence spectra of dimeric carbocyanine dye 1 with G-rich oligonucleotides, single stranded and duplex DNA. Samples contained 2 μM of DNA and dye in 10 mM KCl and 10 mM phosphate buffer (pH 7.0).

Table 1 DNA oligonucleotides used

Sl no.	Name	Sequence
1	G3T-4	5′-GGGTGGGTGGGTGGGT-3′
2	G2T2-4	5′-GGTTGGTTGGTTGGTT-3′
3	G4T4-2	5′-GGGGTTTTGGGGTTTT-3′
4	G3T3-2	5′-GGGTTTGGGTTT-3′
6	G3TG3TG3TG3	5′-GGGTGGGTGGGTGGG-3′
7	G3T2A-3	5′-GGGTTAGGGTTAGGGTTA-3′
8	c-Myc	5′-TGGGGAGGGTGGGGAGGGT-3′
9	SS1	5′-AGGCCACCTTAA-3′
10	D	5′-GAACCACTGAAAGATCTA-3′/5′-TAGATCTTTCAGTGGTTC-3′

---

Fig. 3 Comparative UV-visible absorbance spectra of dimeric carbocyanine dye 1 with G-rich oligonucleotides and duplex DNA. Samples contained 2 μM of DNA and dye in 10 mM KCl and 10 mM phosphate buffer (pH 7.0).

The fluorescence behaviour of 1 in presence of the G-rich oligonucleotides G2T2-4, G3T-4 and G4T4-2 is interesting in this regard. As shown in Fig. 2 the fluorescence of 1 in with G3T-4 is nearly 4–5 fold higher compared to that with G2T2-4 or G4T4-2. These sequences have the same total length but different numbers of contiguous guanines that enable dissimilar quadruplex conformations to be formed. Taken together, these experiments provide an initial indication of the ability of 1 to selectively interact with G-rich oligonucleotides. Atomic Force Microscopy (AFM) reveals a distinctive 10-fold decrease in the size of aggregates formed by 1 in presence of G3T-4 compared to that formed by the dye alone (see Fig. S1†). The more dispersed nature of 1 in presence of G3T-4 follows from the formation of DNA-bound species. In contrast, the AFM images of 1 alone clearly indicate the presence of larger aggregates in close proximity to one another.

CD spectroscopy reveals the different quadruplex folding topologies of the G-rich oligonucleotides in presence of K⁺ (see ESI, Fig. S2†).³⁸ The observed CD maxima and minima of G2T2-4 and G3T-4 correlate with the formation of antiparallel and parallel quadruplexes, respectively. G4T4-2 appears to exist as a mixed quadruplex. The topology and strength of the quadruplex are likely to play a syncretic role in influencing interaction with a small molecule such as 1. Thus, factors that affect the conformation and/or stability of the quadruplex are also expected to carry their imprint on the fluorescence behaviour of 1. A simple change in the monovalent ion in the medium from K⁺ to Na⁺ validates this hypothesis. Change in monovalent ion does not alter conformation of the quadruplexes (see Fig. S3†). The behaviour of 1 with various oligonucleotides is similar in presence of Na⁺ as compared to K⁺, and the fluorescence of 1 is still greatest with G3T-4 (see Fig. S4†).

The interplay of sequence traits such as the number of contiguous guanines and spacer or loop length with the identity of monovalent ion is known to strongly influence quadruplex conformation and stability. We examined this interplay with regards to the ability of 1 to reflect the emergent quadruplex conformation and stability. Addition of 1 to fixed concentrations of the quadruplexes in presence of Na⁺ or K⁺, and measurement of the corresponding fluorescence intensities, permits calculation of binding affinities³⁹ (see Fig. 4, S5 and Table S1†).

Fig. 4 Fluorescence enhancement of dimeric carbocyanine dye 1 upon addition to G3T-4 in presence of K⁺ and Na⁺. Samples contained 2 μM DNA in 10 mM phosphate buffer (pH 7.0).

Two aspects are worth noting from the fluorescence titration experiments. First, Na⁺ rather than K⁺ is found to favor interactions of 1 with the quadruplex formed by G3T-4. The interaction of 1 with G4T4-2 and G2T2-4 are not as affected by the change in the monovalent ion. Second, the binding affinity of 1 differs by nearly an order of magnitude between the parallel and antiparallel quadruplexes. The preference of 1 for parallel quadruplexes is further evident from the differences in binding behaviour between G3T-4 and G3T3-2. While these two oligonucleotides have the same number of contiguous guanines, their spacer lengths are different leading to different conformations of the cognate quadruplexes. In particular, G3T3-2 folds into an antiparallel structure and exhibits weaker binding affinity with 1 in comparison to G3T-4. Na⁺ may be enhancing binding of 1 with the quadruplexes based on its looser fit in the quadruplex cavity^40,41 leading to better intercalation by the dye. The weaker interaction of 1 with the antiparallel quadruplex is possibly indicative of an end-stacking mechanism of binding.

Parallel quadruplexes may possess propeller and lateral loops but do not have loops crossing over the terminal G-tetrads. The presence of such loops in the quadruplexes formed by G2T2-4, G4T4-2 and G3T3-2 possibly imposes steric barriers to effective end-stacking.

We investigated the stoichiometry of binding of 1 with the quadruplexes in order to shed more light on the binding interactions. Fluorescence spectra were measured for samples bearing a range of dye : quadruplex ratios but with constant total amount of the two. The inflection point of such a continuous variations experiment corresponds to the formation of greatest amount of dye : DNA complex. As shown in Fig. 5, for the interaction of 1 with G3T-4 (in K⁺), inflection is observed at a mole fraction of 0.8 which corresponds to one dye per quadruplex molecule. Similar experiments on 1 with G4T4-2 or G2T2-4 exhibit lack of sharp inflection points and indicate a combination of binding stoichiometries (see Fig. S6†). Interestingly, the binding stoichiometry of 1 for G3T-4 approaches 2 dyes per quadruplex in presence of Na⁺ (data not shown). The distinctive binding stoichiometry of 1 with G3T-4 compared to the other quadruplexes is a manifestation of selective binding for the parallel topology. In particular, the parallel conformation provides greater effective recognition surface for interaction both due to the greater number of G-tetrads as well as less steric barriers in terminal section of the structure. Also, since 1 comprises two chromophoric moieties, end-stacking of only one of these on the antiparallel quadruplexes without substantive participation of the other, may also result in variable stoichiometries of binding. The interaction of dye 1 with the parallel quadruplexes was further studied through interaction with G3TG3TG3TG3. This oligo forms parallel quadruplex with the same number of G-quartets as G3T-4 but ends with a terminal guanine rather than thymine and is expected to provide a tighter stacking surface at both ends for the dye. 1 exhibits two-fold greater fluorescence with G3TG3TG3TG3 compared to that with G3T-4 (data not shown).

Fig. 5 Job plot of dye 1 with G3T-4 in K⁺.

While, the continuous variations experiment of 1 with G3TG3TG3TG3 lacks a sharp inflection point it points towards a 1 : 2 (or 2 : 4) binding stoichiometry which conforms to the end-stacking model of binding (see Fig. S7†). To further substantiate the end-stacking mode of binding, FRET experiments were performed involving 1 and 5′ or 3′-Cy5 labeled G3T-4. The significantly greater FRET efficiency between 1 and 5′-Cy5-G3T-4 in contrast to that between 1 and 3′-Cy5-G3T-4 is clearly indicative of the end-stacking of 1 on the quadruplexes (see Fig. S8†).

The effects of dye binding upon the quadruplex conformation and stability were monitored by CD spectroscopy and CD melting experiments. As shown in Fig. 6, 1 was not found to alter the structures of the quadruplexes upon binding. This result is promising towards application of 1 as a reporter of quadruplex structure without infringing on the ability of the latter to interact with other agents.

Fig. 6 CD spectral characterization of G3T-4 before and after dye binding. Samples contained 2 μM DNA and 2 μM dye 1 with 10 mM KCl in 10 mM phosphate buffer (pH 7.0).

We examined the ability of dye 1 to recognize quadruplexes formed by more complex and physiologically relevant sequences. In this regard, the interaction of 1 with the short human telomeric oligonucleotides (G3T2A-3, see Table 1) and separately with the promoter sequence of c-Myc oncogene (c-Myc) leads to fluorescence enhancement of the dye that is comparable in magnitude to that obtained in presence of G3T-4 (see Fig. S9†). The fluorescence enhancement of 1 suggests the presence of parallel quadruplex motifs in these DNA molecules (vide supra) that is in conformity with their known structural characteristics in solution.^42,43 The preference of 1 for parallel quadruplexes was further examined through a series of G3-bearing oligonucleotides (Table S3†). Without exception, the sequence traits and monovalent ion that promote parallel quadruplex formation elicit a superior fluorescence response from 1 (Fig. S10 and S11†).

Conclusions

These results highlight the ability of the dimeric carbocyanine dye 1 to bind specific quadruplex structures with high affinity, as compared to mixed sequence or double-stranded DNA molecules. Dye 1 prefers interaction with parallel quadruplexes and binds via an end-stacking mode.

The interaction of 1 with quadruplexes appears to be contingent on an optimum configuration of number of G-quartets, the spacer length and identity of monovalent ion. The interplay of these factors yields quadruplexes possessing ideal conformation and recognition surface for 1 to interact with. The quadruplex structure is preserved upon dye binding. Quadruplex recognition by small molecule dyes has mostly relied on effective stacking interactions between the participants.²³ While this strategy has enabled reasonable power of distinction between duplexes and quadruplexes, the ability to distinguish between different types of quadruplexes remains challenging. Our results expand the paradigm of G-quadruplex recognition and point towards a fluorescence-based readout of quadruplex conformation and stability. Further studies of the nuances surrounding sequence and structure of quadruplexes with respect to binding by 1 are currently underway in our laboratory and would elaborate on the scope of use of dimeric carbocyanine dyes in this format.

Acknowledgements

B. D. would like to thank the Center for Biomedical Engineering, IIT Gandhinagar for financial support of this work. Experimental assistance provided by Kshiti Patel is gratefully acknowledged.

Notes and references

1. B. A. Armitage, Top. Heterocycl. Chem., 2008, 14, 11.
2. S. Dash, M. Panigrahi, S. Baliyarsingh, P. K. Behera, S. Patel and B. K. Mishra, Curr. Org. Chem., 2011, 15, 2673.
3. H. S. Rye, S. Yue, D. E. Wemmer, M. A. Quesada, R. P. Haugland, R. A. Mathies and A. N. Glazer, Nucleic Acids Res., 1992, 20, 2803–2812.
4. D. Marie, D. Vaulot and F. Partensky, Appl. Environ. Microbiol., 1996, 62, 1649–1655.
5. S. Gurrieri, K. S. Wells, I. D. Johnson and C. Bustamante, Anal. Biochem., 1997, 249, 44–53.
6. A. I. Ahmad and J. B. Ghasemi, Anal. Bioanal. Chem., 2007, 389, 983–988.
7. T. G. Deligeorgiev, S. Kaloyanova and J. J. Vaquero, Recent Pat. Mater. Sci., 2009, 2, 1–26.
8. B. A. Armitage, Organic Photochemistry and Photophysics, 2006, pp. 255–287.
9. A. L. Benvin, Y. Creegar, G. W. Fisher, B. Ballou, A. S. Waggoner and B. A. Armitage, J. Am. Chem. Soc., 2006, 129, 2025–2034.
10. J. L. Seifert, R. E. Connor, S. A. Kushon, M. Wang and B. A. Armitage, J. Am. Chem. Soc., 1999, 121, 2987–2995.
11. K. C. Hannah and B. A. Armitage, Acc. Chem. Res., 2004, 37, 845–853.
12. R. Cao, C. F. Venezia and B. A. Armitage, J. Biomol. Struct. Dyn., 2001, 18, 844–856.
13. B. R. Renikuntla and B. A. Armitage, Langmuir, 2005, 21, 5362–5366.
14. A. N. Glazer and H. S. Rye, Nature, 1992, 359, 859–861.
15. G. T. Hirons, J. J. Fawcett and H. A. Crissman, Cytometry, 1994, 15, 129–140.
16. A. Larsson, C. Carlsson, M. Jonsson and B. Albinsson, J. Am. Chem. Soc., 1994, 116, 8459–8465.
17. K. Gunther, M. Mertig and R. Seidel, Nucleic Acids Res., 2010, 38, 6526–6532.
18. W. Gai, Q. Yang, J. Xiang, W. Jiang, Q. Li, H. Sun, A. Guan, Q. Shang, H. Zhang and Y. Tang, Nucleic Acids Res., 2013, 41, 2709–2722.
19. H. Ihmels and L. Thomas, Org. Biomol. Chem., 2013, 11, 480–487.
20. J. L. Huppert and S. Balasubramanian, Nucleic Acids Res., 2005, 33, 2908–2916.
21. P. Rawal, V. B. R. Kummarasetti, J. Ravindran, N. Kumar, K. Halder, R. Sharma, M. Mukerji, S. K. Das and S. Chowdhury, Genome Res., 2006, 16, 644–655.
22. N. W. Luedtke, Chimia, 2009, 63, 134–139.
23. D. Yang and K. Okamoto, Future Med. Chem., 2010, 2, 619–646.
24. J. Cuesta, M. A. Read and S. Neidle, Mini-Rev. Med. Chem., 2003, 3, 11.
25. J. H. Tan, L. Q. Gu and J. Y. Wu, Mini-Rev. Med. Chem., 2008, 8, 1163.
26. C. L. Grand, H. Y. Han, R. M. Munoz, S. Weitman, D. D. Von Hoff, L. H. Hurley and D. J. Bearss, Mol. Cancer Ther., 2002, 1, 565.
27. J. Ren and J. B. Chaires, Biochemistry, 1999, 38, 16067.
28. S. Ghosh Datta, C. Reynolds, Y. K. Goud and B. Datta, J. Biomol. Struct. Dyn., 2014, 32, 1155–1163.
29. L. Yu, Q. Yang, J. Xiang and Y. Tang, Anal. Methods, 2015, 7, 5483–5489.
30. R. Nanjunda, E. A. Owens, L. Mickelson, T. L. Dost, E. M. Stroeva, H. T. Huynh, M. W. Germann, M. M. Henary and W. D. Wilson, Molecules, 2013, 18, 13588–13607.
31. L. Yu, Q. Yang, J. Xiang, H. Sun, L. Wang, Q. Li, A. Guan and Y. Tang, Analyst, 2015, 140, 1637–1646.
32. Q. Yang, J. Xiang, S. Yang, Q. Li, Q. Zhou, A. Guan, X. Zhang, H. Zhang, Y. Tang and G. Xu, Nucleic Acids Res., 2010, 38, 1022–1033.
33. R. F. Khairutdinov and N. Serpone, J. Phys. Chem. B, 1997, 101, 2602–2610.
34. M. Wang, G. L. Silva and B. A. Armitage, J. Am. Chem. Soc., 2000, 122, 9977–9986.
35. M. Kasha, Physical Processes in Radiation Biology, Academic Press, New York, 1964.
36. A. S. Davydov, Theory of Molecular Excitons, Plenum Press, New York, 1971.
37. W. West and S. Pearce, J. Phys. Chem., 1965, 69, 1894–1903.
38. P. Balagurumoorthy, S. K. Brahmachari, D. Mohanty, M. Bansal and V. Sasisekharan, Nucleic Acids Res., 1992, 20, 4061–4067.
39. M. J. Joseph, J. C. Taylor, L. B. McGown, J. B. Pitner and C. P. Linn, Biospectroscopy, 1996, 2, 173–183.
40. M. S. Kaucher, W. A. Harrell Jr and J. T. Davis, J. Am. Chem. Soc., 2006, 128, 38–39.
41. N. H. Campbell and G. N. Parkinson, Methods, 2007, 43, 252–263.
42. N. Zhang, A. T. Phan and D. J. Patel, J. Am. Chem. Soc., 2005, 127, 17277–17285.
43. R. I. Mathad, E. Hatzakis, J. Dai and D. Yang, Nucleic Acids Res., 2011, 39, 9023–9033.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra05474a

---