G-Quadruplex and i-motif Formation by Repeat Patterns

CD spectroscopy was employed to study if the repeat patterns are potential DNA-G-quadruplex forming sequences in the presence of K⁺, Na⁺ and Li⁺ cations in vitro. Stabilization of G-quadruplexes by monovalent cations is dependent on the nature of the cation, in general the stabilizing effect of K⁺ is greater than that of Na⁺, while Li⁺ shows only poor effects (44). Both the minimal motif needed to form a G-quadruplex consisting only of the four G-tracts and three loop regions d[(GGGAATC)3GGG] as well as the extended repeat motif 5’-d[(GGGAATC)4] were analyzed. In case of the Ana sequences different G-quadruplex patterns are possible with the fourth guanine either being part of the loop sequence d[(GGGACTG)3GGG] or being located in the G-tract d[(GGGGACT)3GGGG]. Therefore both oligonucleotide types with differently long G-tracts in addition to the repeat type oligonucleotide were studied. All samples were prepared with 5 µM oligonucleotide in 10 mM Tris-HCl (pH 7.5) with the respective salt concentration as indicated. All CD spectra were measured from 220-320 nm at 20°C. Different G-quadruplex structures can be distinguished according to their signature in CD, a typical spectrum of an antiparallel quadruplex shows a minimum at 260 nm and a maximum at 290 nm, while a quadruplex with parallel strand orientation shows a minimum at 240 nm and a maximum at 260 nm (113). In addition mixtures of both structures and hybrid structures with both parallel and antiparallel strand orientations can be detected. Schemes depicting the different quadruplex topologies are shown in Figure 5 (Introduction).

For the G-rich motif from Xcc d[(GGGAATC)3GGG] CD spectra in solution with KCl showed a minimum in ellipticity at 240-250 nm, a shoulder at 270 nm and a maximum at 290 nm indicative for a (3+1) hybrid structure, in a which the G-quadruplex is formed from three parallel and one antiparallel oriented strand (Figure 31A). A similar but less pronounced spectra is present for the repeat motif d[(GGGAATC)4] in the presence of K⁺ (Figure 31B). CD spectra in the presence of NaCl showed minor changes in ellipticity compared to the unfolded state, however no specific quadruplex structure can be asigned. Folding in the presence of LiCl did not show significant variation from the unfolded state (Figure 31A and B). Additional spectra in the presence of 50 mM KCl are shown in Figure 53 and did not vary from the spectra observed under higher salt conditions.

Possible quadruplex forming oligonucleotides from Ana showed clear formation of an antiparallel structure in the presence of KCl for d[(GGGGACT)3GGGG] (Figure 31C) and d[(GGGGATT)3GGGG]

(Figure 31D) with a strong maximum at 290 nm and a minimum at 260 nm, while NaCl had a weaker effect. Spectra in the presence of LiCl are equal to the unfolded state. Peaks at 290 nm are also present in the spectra of d[(GGGACTG)3GGG] (Figure 31E) and d[(GGGATTG)3GGG] (Figure 31F) in solution with KCl, however the effect is less pronounced.

Figure 31: CD spectra of Repeat-Derived G-Quadruplex Forming Sequences from Xcc, Xac and Ana A: d[(GGGAATC)3GGG], B: d[(GGGAATC)4], C: d[(GGGGACT)3GGGG], D: d[(GGGGAAT)3GGGG], E: d[(GGGACTG)3GGG], F: d[(GGGATTG)3GGG], G: d[(GGGGACT)4] and H: d[(GGGGATT)4]. CD spectra recorded from 220 to 320 nm of 5 μM oligonucleotide in 10 mM Tris-HCl (pH 7.5) in the presence of 100 mM LiCl (green), 100 mM NaCl (blue), 100 mM KCl (red) or tris buffer only (gray).

For these oligo types four guanines are present in the second and third G-tract, which enables formation of a variety of G-quadruplex structures with 3 guanines in the G-tract, spectra of the different structures formed would then overlap in CD. Again NaCl did not lead to equally pronounced quadruplex formation as KCl and spectra in the presence of LiCl were equal to the unfolded state. In addition we also measured spectra of the repeat motifs d[(GGGGACT)4] and d[(GGGGATT)4], in these cases the spectra of an antiparallel quadruplex were even more pronounced with KCl than for d[(GGGGACT)3GGGG] and d[(GGGGATT)3GGGG]. Presence of NaCl also induced formation of an antiparallel species, while LiCl had no or little effect (Figure 31G and H). Additional spectra in the presence of 50 mM KCl are shown in Figure 53 and did not vary from the spectra observed under higher salt conditions.

Figure 32: CD Spectra of (GGGAATC)³GGG Variants With G to T Mutations in the G-Tract

CD spectra were recorded from 220 to 320 nm of 5 μM oligonucleotide in 10 mM Tris-HCl (pH 7.5) in the presence of 100 mM LiCl (green), 100 mM NaCl (blue), 100 mM KCl (red) or tris buffer only (gray). Oligonucleotide sequences are shown in graphs. G to T muations for G-tract mutations are shown in orange. A: d[GTGAATC(GGGAATC)2GGG].

B: d[GGGAATCGTGAATCGGGAATCGGG]. C: d[(GGGAATC)2GTGAATCGGG]. D: d[(GGGAATC)3GTG].

In a few cases variation of the second position in the G-tract had been noticed for some repeat units. On the one hand any nucleotide other than G at this position may disrupt G-quadruplex formation, however the formation of a structure with two tetrads and a looped out nucleotide in the G-tract may also be possible. To test this, spectra of d[(GGGAATC)3GGG] variants carrying G to T mutations at the second position in the G-tract were recorded. Spectra did not show the typical minima and maxima indicating quadruplex formation in the presence of LiCl, NaCl or KCl, with the exception of d[GTGAATC(GGGAATC)2GGG] where minor stabilization of a structure in the presence of salts was detected (Figure 32).

To gain an insight whether these structures are stable enough to potentially form in vivo thermodynamic stability of the structures formed in the presence of KCl and NaCl was determined by thermal denaturation experiments. Folded oligos were denatured under the same buffer conditions as used for CD measurements by a temperature gradient from 20°C to 100°C while monitoring ellipticity at 290 nm. Melting temperatures T1/2 are listed in Table 7, melting profiles are shown in Figure 54. We determined moderate melting temperatures T1/2 of 50.4°C for the Xcc quadruplex d[(GGGAATC)3GGG] in the presence of 100 mM KCl. Under the same conditions the repeat motif d[(GGGAATC)4] was less stable with T1/2 40.1°C. In both cases T1/2 determined at 270 nm and 290 nm are almost identical suggesting formation of one quadruplex species only and not a mixture of structures with different topologies. In contrast all sequences from Ana proofed to be much more stable than the Xcc quadruplex; in fact species with G-tracts comprising four

guanines d[(GGGGACT)3GGGG], d[(GGGGATT)3GGGG] and d[(GGGGACT)4] could not be fully denatured in presence of KCl and showed T1/2 >95°C. For d[(GGGGACT)3GGGG] and d[(GGGGATT)3GGGG] we determined T1/2 of 92.6°C and 88.6°C, respectively, for 50 mM KCl, Quadruplexes with G-tracts made up of three guanines d[(GGGACTG)3GGG] and d[(GGGATTG)3GGG] and were less stable with 100 mM KCl showing T1/2 of 76.8°C and 74.6°C, respectively. Although showing higher ellipcitity in CD, we again determined that the repeat type motifs d[(GGGGACT)4] and d[(GGGGATT)4] had slightly lower T1/2 than the respective minimal quadruplex forming motifs lacking additional bases at the 3’ end. In all cases structures folded in the presence of 100 mM NaCl were less stable than their K⁺ stabilized counterparts.

Table 7: Melting Temperatures of G-Quadruplex Structures

Temperature of the half-maximal decay of ellipticity T1/2 was determined at 290 nm unless otherwise indicated. CD spectra for all oligonucleotides in the presence of 50 mM KCl are shown in Figure 53A. Errors are the errors of the fit.

Sequence (5‘-3‘) T1/2 [°C]

When assessing the folding behavior of the quadruplex forming oligonucleotide from Xcc in the presence of its complementary strand d[(CCCGATT)3CCC], CD spectra showed a regular duplex spectrum in the presence of KCl and (Figure 53B), folding of the G-quadruplex structure did not prevail over duplex formation.

Presence of G-rich repeat patterns inevitably is tied to the presence of a C-rich pattern in the complementary strand, it may therefore also be the C-rich repeats which exhibit a potential biological function. The so-called i-motif structure is formed from C-rich oligonucleotides at mild acidic conditions, which enables the formation of hemiprotonated cytosine-cytosine+ base pairs (see scheme in Figure 9, Chapter 1.4.1) (62). Although formation of the i-motif is favored at lower pH, some sequences are able to stably fold i-motif structures even at neutral pH (150). It might form under special environmental conditions, especially under negative supercoiling conditions (62,104). i-motif formation can be studied with CD spectroscopy showing a characteristic minimum at about 260 nm and a maximum at around 290 nm (268). We determined the folding behavior of the complementary C-rich repeat strands in presence of different cations as well as at decreasing pH. For the C-rich oligonucleotides derived from Xcc CD spectra showed a slightly shifted spectrum with a minimum at 240 nm, shoulder at 260-270 nm and maximum at 280 nm

at pH 4.5, similar to the spectrum observed for the G-rich oligonucleotides (Figure 33A and B).

Lowering the pH caused i-motif formation in all C-rich oligonucleotides derived from Ana (Figure 33C-H). Remarkably all of the observed structures persisted even at higher pH of 6.5.

Figure 33: pH-Dependent i-Motif Formation of C-rich Repeat Oligonucleotides

CD spectra recorded from 220 to 320 nm of 5 μM oligonucleotide in 10 mM Tris-HCl adjusted to different pH: pH 7.5 (gray), pH 6.5 (green), pH 5.5 (blue) and pH 4.5 (red) for A: d[(CCCGATT)3CCC], B: d[(GATTCCC)4], C: d[(CCCCAGT)3CCCC], D: d[(CCCCAAT)3CCCC], E: d[(CCCCAGT)3CCC], F: d[(CCCCAAT)3CCC], G: d[(AGTCCCC)4] and H: d[(AATCCCC)4].

The thermodynamic stability of the structures formed under acidic conditions was also assessed, (Figure 34, melting profiles in Figure 55). Due to the temperature dependent pH-change of tris buffer all structures were melted in 10 mM sodium acetate buffer pH 4.5 or pH 6.5 with and without additional 100 mM NaCl or KCl. Melting profiles are shown in Figure 55 the respective CD spectra are shown in Figure 34. At pH 4.5 all structures are fairly stable with T1/2 ranging between 60-72°C at pH 4.5 (Table 8). Variations in ionic strength of the buffer may change the stability of i-motif structures, i-motifs have been reported to be destabilized by increased ion concentrations (269,270).

Table 8: Melting Temperatures of Structures Formed by C-rich Repeat Oligonucleotides

Temperature of the half-maximal decay of ellipticity T1/2 was determined at 285 nm unless otherwise indicated. Errors are the errors of the fit.

Sequence T1/2 [°C]

pH 4.5 10 mM Na⁺

T1/2[°C]

pH 4.5 110 mM Na⁺

T1/2[°C]

pH 4.5 10 mM Na⁺ 100 mM K⁺

T1/2[°C]

pH 6.5 10 mM Na⁺

T1/2 [°C]

pH 6.5 110 mM Na⁺ d[(CCCGATT)3CCC] (280 nm) 64.0 ± 0.3 72.6 ± 0.3 72.1 ± 0.3 44.9 ± 0.2 37.0 ± 0.1 d[(GATTCCC)4](280 nm) 60.9 ± 0.3 73.3 ± 0.3 73.4 ± 0.3 45.4 ± 0.1 40.5 ± 0.1 d[(CCCCAGT)3CCCC] 70.0 ± 0.3 74.7 ± 0.2 74.5 ± 0.3 41.2 ± 0.1 30.1 ± 0.1 d[(CCCCAAT)3CCCC] 71.8 ± 0.3 76.4 ± 0.2 76.7 ± 0.2 43.7 ± 0.1 33.9 ± 0.2 d[(CCCCAGT)3CCC] 63.5 ± 0.2 67.7 ± 0.2 67.9 ± 0.2 36.4 ± 0.1 26.1 ± 0.5 d[(CCCCAAT)3CCC] 65.4 ± 0.2 69.7 ± 0.2 69.6 ± 0.3 36.0 ± 0.1 24.9 ± 1.1 d[(AGTCCCC)4] 67.7 ± 0.4 74.7 ± 0.5 73.8 ± 0.4 63.7 ± 0.3 30.7 ± 0.2 d[(AATCCCC)4] 70.3 ± 0.5 78.4 ± 0.3 77.7 ± 0.2 48.3 ± 0.1 37.1 ± 0.1

Figure 34: CD Spectra of C-rich Repeat Oligonucleotides at pH 4.5 and pH 6.5

CD spectra recorded from 220 to 320 nm of 5 μM oligonucleotide in 10 mM sodium acetate buffer pH 4.5 (red), pH 4.5 with additional 100 mM NaCl (light blue), pH 4.5 with additional 100 mM KCl (blue), pH 6.5 (orange), pH 6.5 with additional 100 mM NaCl (green). A: d[(CCCGATT)3CCC], B: d[(GATTCCC)4], C: d[(CCCCAGT)3CCCC], D: d[(CCCCAAT)3CCCC], E: d[(CCCCAGT)3CCC], F: d[(CCCCAAT)3CCC], G: d[(AGTCCCC)4] and H: d[(AATCCCC)4].

However, addition of further 100 mM NaCl or KCl did not disturb i-motif formation. i-motif signatures were readily detectable in CD and T1/2 did not decrease upon salt addition at pH 4.5.

Contrarily we detected an increase in T1/2 of 4-12°C (Table 48). Raising pH to pH 6.5 lead to a destabilization of the formed structures with T1/2 dropping by 15-29°C in comparison to the T1/2

determined at pH 4.5, except for d(AGTCCCC)4, which showed a weaker decrease of only 4°C.

Increasing ionic strength at pH 6.5 did no longer have stabilizing effect and T1/2 decreased further.

As expected i-motif formation was neither detected at pH 7.5 nor influenced by further addition of 100 mM LiCl, NaCl or KCl (Figure 56). For d[(CCCGATT)3CCC] and d[(GATTCCC)4] presence of cations seemed to even destabilize the undefined structure formed in tris buffer only at pH 7.5.

In summary, we could detect changes in CD spectroscopy that indicate G-quadruplex formation in vitro for all G-rich oligonucleotides tested. In case of the Xcc quadruplex sequence we assume that a (3+1) hybrid structure is adopted in the presence of KCl, while all sequences derived from Ana showed spectra characteristic for antiparallel quadruplexes. All over presence of K⁺ cations had a greater stabilizing effect while quadruplex formation assisted by Na⁺ was less pronounced and structures showed lower thermodynamic stability. K⁺ has been reported to be the major cation in the bacterial cell, cytosolic concentrations of about 200 mM were determined for E. coli (234). A concentration of 100 mM K⁺ therefore represents a concentration likely to be achieved in a cellular environment to stabilize potential G-quadruplexes. In fact higher intracellular concentrations of K⁺ up to 500 mM were measured under osmotic shock or salt stress conditions in bacteria (235).

In addition changes in ellipticity that suggest i-motif formation was detected for the complementary C-rich repeat sequences even under only mildly acidic conditions. The structures were highly stable under acidic pH, but showed minor stability at pH 6.5, especially under high salt conditions. CD spectroscopy alone is no definite proof for adoption of a G-quadruplex structure and further analysis concerning the individual conformations need to be carried out.

However characteristic spectral changes and enhanced thermodynamic stability were indeed observed under conditions favoring G-quadruplex formation. Furthermore no structural changes could be observed upon introduction of G to T mutations for the Xcc derived oligonucleotides.

Taken together this strongly suggests that the adopted structures in presence of K⁺ are G-quadruplexes.