The CD spectra were obtained using a Jasco J-810 spectropolarimeter equipped with a six cell, programmable Peltier junction temperature controlled cell holder. The DNA samples used in the CD experiments were at a concentration of 1.1 × 10⁻⁵ M. The buffers used for the CD experiments were 20 mM phosphate, 140 mM total Na and 0.1 mM EDTA at pH 7.0; 2.5 mM phosphate, 5 mM total Na and 0.1 mM EDTA at pH 7.0. The KCl additions increased the volume of the samples <10%. The DNA samples were annealed by heating to 363 K and then cooled to room temperature over a period of 8 h before obtaining the CD spectra. The CD measurements used a 0.2 cm path cell and were carried out at 293 K. The CD spectra were averaged over three scans and the data were obtained with a 1 nm slit width from 350 to 200 nm at 0.1 nm intervals. The CD spectrum of the buffer solution was subtracted from that of the sample containing DNA.

The solution conditions for the four reference DNAs all contained 0.1 mM EDTA at pH 7.0. The d(GGTTGGTGTGGTTGG) solution also contained 100 mM sodium and 140 mM potassium. The d(GGGGTTTTGGGG) solution contained 140 mM sodium. The d(TGGGGT) solution contained 140 mM sodium and 100 mM potassium. The d(TAGGGUTAGGGT) solution contained 5 mM sodium and 100 mM potassium.

The NMR spectra were obtained using a Varian Unity Inova 500 spectrometer operating at a magnetic field of 11.75 T. The spectra were acquired with the samples at 298 K with a repetition rate of 0.17 s. The water solvent signal suppression was accomplished by using the watergate pulse sequence. The samples were prepared in 95% H₂O/5% ²H₂O and were annealed for ∼10 min at 373 K followed by slow cooling to room temperature. The DNA concentrations were 0.5–1.0 mM.

The DNAs were obtained from Integrated DNA Technologies, Inc. (Coralville, Iowa). The DNA samples were supplied as HPLC-purified samples and desalted by ethanol precipitation before use. The concentration of each DNA sample was determined from the absorbance value at 260 nm at 293 K of a sample prepared in distilled H₂O by using the nearest neighbor extinction coefficients. The hypochromicity of each DNA was determined by melting and the hypochromicities were found to be in the range of 9–14 %. The extinction coefficients of the vertebrate repeat DNA samples studied here and the naming convention used are listed below.

The CD spectra of d(TTAGGGTTAGGG), Vet T2, in the presence of 5 and 140 mM sodium are shown in Figure 3. In the presence of 5 mM sodium, the CD spectrum is similar to that of a single-stranded DNA with a positive band near 250 nm. The proton NMR spectrum of this DNA, shown in Supplementary Material, indicates that this DNA is disordered in 5 mM sodium. When the DNA is in 140 mM sodium the spectrum is characteristic of an anti-parallel structure. The spectrum in 140 mM sodium does not contain a large negative peak near 260 nm and this may be owing to the presence of a mixture of both four-stranded parallel and anti-parallel structures. The NMR spectrum (data not shown) of the sample in 140 mM sodium indicates that quartets and also multiple conformers of the DNA are present.

The CD spectrum changes to that of an external loop quadruplex structure in the presence of 100 mM potassium with either 5 or 140 mM sodium as shown in the results in Figure 3. The CD spectrum is essentially the same with 5 or 140 mM sodium when 100 mM potassium is present. There are small changes in the CD spectrum upon the addition of 10 mM potassium to the DNA in the presence of 5 or 140 mM sodium. The Supplementary Material contains the NMR spectrum of this DNA in the presence of 100 mM potassium. The NMR results are consistent with an external loop structure formed by a symmetric dimer.

The CD and NMR results for d(TTAGGGTTAGGG) indicate that in low salt conditions, 5 mM sodium with up to 10 mM potassium, the DNA is primarily in a disordered, single-strand state. In the presence of 140 mM sodium the DNA has a CD spectrum characteristic of an anti-parallel structure. In the presence of 100 mM potassium, the CD spectrum indicates that an external loop parallel-strand structure is present with sodium concentrations in the range of 5–140 mM.

The CD spectra of DNAs with 2.5, 3.5 and 4 repeats of the vertebrate telomere sequence have been examined under the same conditions as described above for d(TTAGGGTTAGGG), Vet T2. In the presence of 140 mM sodium these DNAs have CD spectra characteristic of anti-parallel structures as shown in Figures 4 and 5. All of the spectra contain positive bands near 290 nm and negative bands near 260 nm. These results indicate that these vertebrate telomere repeats adopt anti-parallel structures in the presence of 140 mM sodium.

The spectra of the vertebrate telomere repeat DNAs in the presence of 5 mM sodium vary depending on the number of repeats as shown in the results in Figures 4 and 5. The spectra of the Vet 3.5 and Vet 4 are of the anti-parallel type whereas those of the Vet 2.5 and Vet T4 samples appear to indicate that more than one structural type is present when the DNAs are in the presence of 5 mM sodium.

The presence of 100 mM potassium induces the formation of structures that have CD spectra characteristic of external loop structures both in the presence of 5 mM and in the presence of 140 mM sodium for Vet 2.5, Vet 3.5, Vet 4 and Vet T4, as the results demonstrate in Figures 4 and 5. The changes induced by potassium are analogous to those presented above for Vet T2. The presence of 100 mM potassium induces the formation of structures with CD spectra that are characteristic of external loop structures whether 5 or 140 mM sodium is present.

The CD spectra indicate that for Vet 2.5 10 mM potassium induces more change when in the presence of 140 mM sodium than in the presence of 5 mM sodium whereas the opposite is the case for Vet 4, Vet T4, Vet 3.5 and Vet T2. The potassium effect for each of these vertebrate repeat DNAs is not an ionic strength effect since 140 mM sodium gives distinctly different CD spectra than does the combination of 100 mM potassium and 5 mM sodium.

The ratio of the intensity of the maxima near 260 and 290 nm varies between the vertebrate telomere repeat DNAs but the overall pattern is the same for each of the DNAs in the presence of 100 mM potassium. The CD results indicate that for this set of vertebrate telomere repeat DNAs the presence of 100 mM potassium favors the formation of external loop structures.

Crystal structures of the external loop quadruplex structures of d(AGGG{TTAGGG}₃) and d(TAGGGTTAGGGT) have been obtained (8,35). The DNAs were crystallized from a solution that contained 50 mM potassium. The initial conditions for the hanging drops contained 100 mM potassium for the d(TAGGGTTAGGGT) crystals and 300 mM potassium for the d{AGGG(TTAGGG)₃} crystals.

The CD spectra of d(GGGTTAGGG) and d(GGGTTTGGG) were compared in order to gain information about the effect of the dA residue. The CD spectra of both samples show similar potassium sensitivity as shown in the Supplementary Material. There are some differences in the intensities of the bands near 290 and 260 nm but the general patterns of the spectral changes are similar in the two cases.

DNAs containing portions of the c-myc promoter sequence have been shown to form a parallel-strand external loop structure in the presence of 100 mM potassium (29,30). The CD spectra of these c-myc promoter DNAs have a positive band at 260 nm but do not exhibit the positive band at 290 nm observed from the vertebrate telomere DNAs (29). The external loops of the vertebrate telomere DNAs contain three residues, d(TTA), while those of the c-myc DNAs contain a single dA residue. This suggests that the 290 nm band is due to the loop residues and the 260 nm band is due to the parallel quartet dG residues. The differences observed between d(GGGTTAGGG) and d(GGGTTTGGG) is also in the relative intensities of the 260 and 290 nm bands.

A comparison of the crystal structures of four-strand parallel and external loop structures indicated that the quartet regions have very similar structures in the two cases (8,35). This is consistent with the quartets of both external loop and four-stranded parallel-strand structures giving rise to the positive 260 nm band and the loop residues at the 290 nm band.

In contrast to the behavior of the DNAs containing vertebrate telomere repeats the CD spectrum of the Oxytricha repeat DNA, d(GGGGTTTTGGGG), is of the anti-parallel type under all the conditions examined, as shown in the results in Figure 6. An NMR-based study showed that the addition of potassium to this DNA alters the loop structure but not the basket type anti-parallel quartet structure (13). The structure of a DNA containing four repeats of the Oxytricha sequence has been shown to form an anti-parallel structure in the presence of 100 mM sodium (48).

The Tetrahymena repeat DNA, d(TGGGGTTGGGGT), has been shown to be a mixture of two anti-parallel structures in the presence of 100 mM sodium (47). The CD data for this DNA shown in Figure 6 are in accordance with the NMR derived structure. The CD spectrum of d(TGGGGTTGGGGT) in 5 mM sodium indicates that some four-stranded parallel structure may be present. These results indicate that high sodium concentration is needed for this DNA to adopt anti-parallel structures. The addition of 100 mM potassium to this DNA, in either 5 or 100 mM sodium, suggests that potassium induces the formation of parallel type structures as indicated by the positive band near 260 nm.

The behavior of the CD spectra of both the Oxytricha and the Tetrahymena repeat DNAs are different from that observed for any of the vertebrate telomere repeats. These results indicate that the Oxytricha and Tetrahymena repeat DNAs form different types of structures in the presence of physiological levels of potassium than do the vertebrate telomere repeats.

The CD spectrum of the thrombin binding aptamer d(GGTTGGTGTGGTTGG) remains characteristic of an anti-parallel structure over the range of potassium concentrations from 1–100 mM. The CD spectrum in the presence of 100 mM potassium is shown in Figure 2 and the CD spectra obtained in the presence of other concentrations of potassium have been reported previously. The thrombin binding aptamer requires ∼1 mM potassium, or a suitable substitute, for the formation of the chair type anti-parallel structure (9,10). This behavior was not observed for any of the telomere repeat DNAs examined here and indicates that the thrombin binding aptamer, similar to the Tetrahymena and Oxytricha repeats, is not a good model system for the quadruplex structures formed by the vertebrate telomere repeat.

The CD spectra of d(GGGTTAGGGTTA), Vet 2, shown in Figure 7 do not show the characteristic pattern of external loop structures when the DNA is in the presence of 100 mM potassium. The CD spectrum of Vet 2 is largely unchanged upon the addition of potassium whether the DNA is in the presence of 5 or 140 mM sodium. The CD spectrum does not appear to be comparable with any of the four reference types.

The proton NMR spectrum of Vet 2 has been obtained in the presence and in the absence of 100 mM potassium and these spectra are included in the Supplementary Material. The imino proton spectrum in the absence of potassium is consistent with a quadruplex structure formed by asymmetric dimers as there are at least 12 dG imino resonances. The imino proton spectrum of the DNA in the presence of 100 mM potassium indicates that quartets are present and that there are also resonances downfield of 13 p.p.m. that are indicative of the presence of A–T base pairs. It is noted that A–T base pairs are consistent with the DNA adopting a four-stranded parallel structure but other quadruplex structural types have been observed to allow A–T pairs to form (49). The structure of d(GGGTTAGGGTTA) in the presence of potassium is under investigation and may be of a novel type.

A prior study of Vet T4 and Vet 3.5 found that the reactivity of the N7 positions of the dA residues are different for the samples in 70 mM sodium when compared with the samples in 70 mM potassium as are the gel mobilities (50). Their gel results showed one predominant band for Vet T4 in 70 mM potassium and for both Vet T4 and Vet 3.5 in 70 mM sodium. We have obtained one predominant band for these and other vertebrate telomere repeat DNAs in 100 mM potassium (data not shown). The CD spectra of Vet T4 and Vet 3.5 that they obtained (50) are quite similar to the ones shown here. External loop structures had not been discovered when this prior study was published and hence CD spectra of these types of structures were not available for comparison. They assigned the CD spectra of the vertebrate telomere repeats obtained in the presence of sodium to anti-parallel structures as is performed in this study.

The reactions of d(AGGG{TTAGGG}₃) with platinum (51) and ¹²⁵I degradation (52) have been studied previously (52). An NMR study of d(AGGG{TTAGGG}₃) showed that in 100 mM sodium the structure is an anti-parallel basket type that has low stability for a quadruplex structure requiring the NMR data to be acquired with the sample at 280 K (53). This DNA exhibited imino proton exchange rates with water on the timescale of seconds, even at 280 K, indicating that the structure has low kinetic stability (53). Other quadruplex DNAs have been shown to have imino proton exchange rates on the timescale of days to weeks for the dG residues in the quartets. When crystallized in the presence of high levels of potassium d(AGGG{TTAGGG}₃) was found to have an external loop type structure (8,35). The NMR spectrum of this DNA in the presence of potassium is not interpretable. The low thermal and kinetic stabilities of d(AGGG{TTAGGG}₃) may reduce the utility of this DNA in chemical-based structural studies.

The results of the platinum reactions d(AGGG{TTAGGG}₃) were taken to indicate that a basket type structure is present in both 50 mM sodium and 50 mM potassium (51). The reaction of a dG residue with platinum requires that the N7 is accessible. If a dG residue is present in a quartet this means that the platinum reaction may be predominantly occurring from a disrupted or partially disrupted state. The platination experiments were carried out under conditions, 310 K, at which the structure of the DNA is likely to be significantly disrupted.

The structure of modified forms of d(AGGG{TTAGGG}₃) that contained extra residues on the 3′ and/or 5′ ends have also been investigated using ¹²⁵I. The decay of ¹²⁵I cleaves the DNA at a rate inversely proportional to the distance from the radiolabel. The ¹²⁵I cleavage reactions were carried out on frozen samples and the results in the presence of sodium are consistent with the NMR anti-parallel basket type quadruplex structure obtained in the presence of sodium. Significant differences in the ¹²⁵I cleavage pattern obtained with 100 mM potassium and 100 mM sodium were observed. The observed cleavage pattern for the modified forms of d(AGGG{TTAGGG}₃) are not consistent with either an anti-parallel basket type structure or an external loop type structure.