Quadruplex melting

Abstract

Melting curves are commonly used to determine the stability of folded nucleic acid structures and their interaction with ligands. This paper describes how the technique can be applied to study the properties of four-stranded nucleic acid structures that are formed by G-rich oligonucleotides. Changes in the absorbance (at 295 nm), circular dichroism (at 260 or 295 nm) or fluorescence of appropriately labelled oligonucleotides, can be used to measure the stability and kinetics of folding. This paper focuses on a fluorescence melting technique, and explains how this can be used to determine the T_m (T_½) of intramolecular quadruplexes and the effects of quadruplex-binding ligands. Quantitative analysis of these melting curves can be used to determine the thermodynamic (ΔH, ΔG, and ΔS) and kinetic (k₁, k₋₁) parameters. The method can also be adapted to investigate the equilibrium between quadruplex and duplex DNA and to explore the selectivity of ligands for one or other structure.

Introduction

G-rich nucleic acid sequences can fold into four-stranded DNA structures that contain stacks of G-quartets [1], [2], [3], [4]. These quadruplexes (tetraplexes) can be formed by the intermolecular association of four DNA molecules, dimerization of sequences that contain two G-tracts, or by the intramolecular folding of a single strand containing four blocks of guanines.

DNA melting studies have been widely employed to investigate the stability of duplex DNA and its interaction with ligands [5], [6], [7]. GC-rich sequences melt at higher temperatures than AT-containing ones and compounds that bind to double-stranded DNA selectively stabilise this form over the single-stranded random coil, thereby increasing the melting temperature. These duplex melting experiments are usually (though not exclusively) performed with high molecular weight fragments (natural DNAs or synthetic polynucleotides) and the DNA melts in a highly cooperative fashion. The melting transitions are usually detected by measuring the change in absorbance at 260 nm, which increases by about 25% on denaturation. Similar studies can be performed with higher order DNA structures (e.g., triplexes or quadruplexes), though the results for triplexes may be difficult to interpret as two or more transitions are present (triplex to duplex and duplex to single strands). DNA quadruplexes show only small changes in absorbance at their UV maximum (260 nm), and so this technique has been less widely employed. However, a greater signal is obtained at 295 nm, at which there is a large decrease in absorbance on melting [8]. Other means of monitoring the melting behaviour have therefore been employed including circular dichroism [9], [10], NMR [8], [11] and fluorescence melting [12], [13], [14], [15], [16] studies. Two different CD spectral signatures have been described for quadruplexes, which are usually attributed to the presence of distinct conformations; a peak around 265 nm is thought to correspond to the parallel-stranded form, in which the nucleotides are arranged all anti, while a positive maximum around 295 nm corresponds to one of several antiparallel forms that contain nucleotides in both the syn and anti conformation [17], [18], [19]. The melting of both these forms can be followed by the changes in the CD spectra at appropriate wavelengths [9], [10], [20], [21], [22]. In contrast to duplex DNA, most studies with quadruplex DNA have employed synthetic oligonucleotides which form intermolecular (tetramolecular and dimeric) or intramolecular structures. Since it is straightforward to add fluorescence reporter groups during oligonucleotide synthesis, the use of fluorescence melting profiles has become a popular method for measuring quadruplex stability and its interaction with ligands.

There is only a small (4%) change in absorbance at 260 nm (the wavelength used for studying duplex melting) on G-quadruplex formation. However, although the absolute absorbance is lower at higher wavelengths the change is greatest around 295 nm, for which there is a 50–80% increase in absorbance on quadruplex formation (i.e., a decrease on melting) [8]. This is therefore the preferred wavelength for obtaining high quality melting data that can be used for determining thermodynamic properties of quadruplexes. Since this transition is inverted relative to that of duplex melting profiles, it provides a good means for comparing quadruplex and duplex formation. The values measured at this wavelength compare well with those determined by other methods, such as CD or NMR [8]. A typical absorbance melting profile is shown in Fig. 1b and c. The difference between the absorbance spectra of the folded and single-stranded forms (thermal difference spectrum) is also characteristic of quadruplex DNA, and is distinct from that of other DNA structures [23].

Since quadruplex DNA structures have distinctive circular dichroism spectra, temperature dependent changes in CD have often been used to determine quadruplex stability [9], [10], [20], [21], [22]. The wavelength used depends on the specific sequence (260 nm for intermolecular parallel complexes), but 295 nm for intramolecular antiparallel quadruplexes. These experiments require sophisticated equipment, which is not available in many laboratories, and use relatively high oligonucleotide concentrations (typically 5 μM or higher). In addition these experiments often do not have good temperature resolution.

In this technique [12], [16], synthetic oligonucleotides are prepared that contain a fluorophore (typically fluorescein) at one end and a fluorescence quencher (dabcyl or methyl red) at the other [12], [24], [25]. The principle of this method is illustrated in Fig. 1a. When the oligonucleotide adopts a folded configuration the reporter groups are close together and the fluorescence is quenched. When the structure melts these groups are separated and there is a large increase in the fluorescence signal. We routinely use this technique to compare the stability of different DNA quadruplexes [26], [27], [28], [29], [30], [31], [32]. A typical fluorescence melting profile is shown in Fig. 1b and c. Some laboratories use fluorescence donor and acceptor pairs (such as fluorescein and TAMRA) so that there is FRET (fluorescence energy transfer) between the donor and acceptor in a distance related fashion [13], [14], [33], [34], [35], [36], [37], [38]. However, it should be noted that, although such FRET oligonucleotides give excellent results, the distance between the two groups in the folded complex is usually too short for FRET and the decrease in donor emission on quadruplex formation is probably the result of collisional quenching rather than FRET. This quenching can however be reduced by adding suitable spacer nucleotides between the quadruplex and the fluorescent probes [39], [40]. Most experiments that use FRET pairs for fluorescence melting only report on changes in the fluorescence donor rather than the acceptor. For intermolecular complexes it is also possible to use oligonucleotides that possess a single fluorescent label which undergoes self-quenching on quadruplex formation [27]. If the quadruplex-forming region is located within a longer sequence the fluorophores and/or quenchers can be placed within the oligonucleotide (instead of at the 5′- and 3′-ends) usually by attachment to the 5-position of T, or by incorporation of dR-FAM as an internal unpaired base [41]. However, these internal modifications can have a greater effect on stability, which needs to be checked. Note that the fluorescent reporter groups are attached to the oligonucleotides by relatively long chemical linkers, and there is no significant difference in quenching efficiency between different folded forms. This technique therefore cannot be used to distinguish between the various quadruplex topologies (parallel and antiparallel).

The fluorescence melting technique has several advantages over conventional UV absorbance studies. First, absorbance changes are not large (typically only 25%), while the fluorescence signal can change by 10-fold or greater. Second, UV absorbance melting is usually a low throughput technique, as most spectrophotometers handle no more than six samples at once. Third, the typical format used for fluorescence melting experiments (using real-time PCR machines as described below) only requires small amounts of material (typically 20 μl of a 0.25 μM solution) in contrast to UV studies that require relatively large volumes (1–3 ml) of a solution with an OD₂₆₀ of at least 0.2 (i.e., a total of about 20 nmole of bases). The lower concentrations will also favour the formation of intramolecular (rather than intermolecular) quadruplexes.

Although fluorescence melting is a convenient and simple technique, it requires the attachment of reporter molecules to the oligonucleotide which may influence quadruplex structure and/or stability. The results should therefore be compared with UV melting studies to assess the effects of the added fluorophores. Fig. 1b and c shows a comparison of fluorescence melting profiles with UV melting data with for labelled and unlabelled oligonucleotides based on the human telomeric repeat sequence. It can be seen that the two methods give very similar profiles, though addition of the fluorophores decreases the T_m by a few degrees. While fluorescent dyes have been shown to have relatively little effect on duplex [42] and triplex [43] stability, studies with fluorescently labelled quadruplex [14] and i-motif [44] forming oligonucleotides show a more pronounced effect on stability. Generally, the addition of fluorescent groups is destabilising, although the observed thermodynamic effects can be dye specific. Whenever possible, CD spectra should be determined to ensure that the added groups have not radically altered the topology, indeed there is one report of an instance in which FAM/TAMRA tagged oligos show differences in the CD spectra [14].