Opposite Effects of Potassium Ions on the Thermal Stability of i-Motif DNA in Different Buffer Systems.

i-motifs are noncanonical DNA structures formed via the stack of intercalating hemi-protonated C+: C base pairs in C-rich DNA strands and play essential roles in the regulation of gene expression. Here, we systematically investigated the impacts of K+ on i-motif DNA folding using different buffer systems. We found that i-motif structures display very different T m values at the same pH and ion strength in different buffer systems. More importantly, K+ disrupts the i-motif formed in the MES and Bis-Tris buffer; however, K+ stabilizes the i-motif in phosphate, citrate, and sodium cacodylate buffers. Next, we selected phosphate buffer and confirmed by single-molecule fluorescence resonance energy transfer that K+ indeed has the stabilizing effect on the folding of i-motif DNA from pH 5.8 to 8.0. Nonetheless, circular dichroism spectra further indicate that the structures formed by i-motif sequences at high K+ concentrations at neutral and alkaline pH are not i-motif but other types of higher-order structures and most likely C-hairpins. We finally proposed the mechanisms of how K+ plays the opposite roles in different buffer systems. The present study may provide new insights into our understanding of the formation and stability of i-motif DNA.

Introduction

Specific DNA sequences in suitable environments can fold into noncanonical secondary structures and function as the regulatory elements in DNA metabolism.[1,2] For instance, the G-rich regions in the genome are able to form G-quadruplexes (G4s) by Hoogsteen bonds.[3,4] Interestingly, the C-rich strand complementary to the G4 sequence can fold into another type of a noncanonical structure named i-motif which was first discovered by Gehring et al. in 1993.[5] It is well known that the i-motif structure is formed by the intercalation of hemi-protonated C+: C base pairs at acid conditions and the existence of i-motif has been confirmed in the human cell nucleus.[6,7] Since the i-motif structure was reported, its biological significance has been continuously demonstrated. For instance, the i-motif in the insulin-linked polymorphic region (ILPR) and c-MYC promoter could repress DNA replication and transcription.[8,9] Moreover, the formation of the i-motif structure in the Bcl2 oncogene promoter can upregulate the Bcl2 gene expression, while the C-hairpins lower the Bcl2 levels.[10,11] Besides the biological functions, i-motif structures also play important roles in studying the energy conversion and material transfer mechanisms in living organisms relying on the sensitive response to pH. Therefore, the development of i-motif-based biosensors, soft materials, and motor devices has surged in these years.[12−14] Exploring the effects of monovalent cations on i-motif folding can allow us to not only have a clearer understanding of the structure but also have potential significance for its applications as those cations may exist inside the cell.

Similar to other nucleic acid structures, the stability of i-motif is affected by several factors including the DNA sequences, pH, temperature, ionic strength, molecular crowding conditions, and so forth. A lot of studies have been focused on the effect of pH on i-motif.[15−17] In contrast, only a few studies reported the effects of ionic strength, and different conclusions have been obtained by different labs. For example, Day et al. have been working on the regulation of i-motif structures by Ag+ and Cu2+ ions in recent years. They suggested that Ag+ could induce the human telomeric C-rich sequence to fold into i-motif at room temperature and physiological pH while Na+ could not.18 In contrast, Cu2+ is capable of altering the conformation of i-motif into C-hairpin structures even at acidic pH,[19] and this process is reversible after the addition of ethylenediaminetetraacetic acid (EDTA).[20] Besides, Kim et al. found that Li+ suppressed the formation of i-motif significantly, but Na+ has the stabilizing effect.[21] Mergny et al. reached a different conclusion from Kim et al. on the role of Na+ ions. They found that increasing the Na+ concentration from 0 to 100 mM produced a destabilization on the i-motif structure; however, higher Na+ concentrations did not lead to further destabilization.[22] The above-mentioned seemly controversial phenomena suggest that further investigations of the ionic effect on the folding of the i-motif DNA are needed.

In this work, we systematically investigated the impacts of the monovalent ion on the stability and folding status of i-motif DNA. K+ was selected as it is ubiquitous inside the cell. By the fluorescence resonance energy transfer (FRET)-melting assay, we discovered that K+ has opposite effects on the thermal stability of i-motif structures in different buffer systems. Specifically, K+ disrupts i-motif structures formed in MES and Bis-Tris; however, K+ stabilizes i-motif structures in PB (phosphate buffer), SSC (saline sodium citrate buffer), and SCB (sodium cacodylate buffer). Thereafter, we predominantly selected PB to further determine the folding status of i-motif DNA owing to its wide buffer range. We employed both circular dichroism (CD) and single-molecule FRET (smFRET) which allows us to accurately record the folding/unfolding of i-motif structures. Our single-molecule level results indicate that the high concentration of K+ indeed promotes the folding of i-motif DNA in PB at pH 5.8, 7.0, and 8.0. However, CD spectra suggest that the increase in thermal stability at pH 7.0 and 8.0 does not represent the formation of the i-motif structures. Instead, the C-hairpins are likely to be formed. We also characterized the influence of the complementary strands on i-motif DNA at different pH and ionic conditions. At acidic pH, the free G4 strands show little effects on i-motif folding; however, at neutral conditions, the formation of duplex DNA is favored regardless of the K+ concentration. Finally, we proposed the mechanisms to explain how K+ plays the opposite roles in modulating i-motif DNA in different buffer systems.

Experimental Materials and Methods

Buffers

In this work, we selected PB (K2HPO4–KH2PO4), MES (C6H13NO4S–NaOH), Bis-Tris (C8H19NO5– HCl), SSC (Na3C6H5O7–C6H8O7), and SCB (C2H12AsNaO5–HCl) buffer systems. As there is inherent K+ in the PB, the concentration of K+ in the following study means the additional K+ we added. For the single-molecule measurement, the folding buffer of i-motif DNA contained 0–750 mM KCl in 50 mM PB, pH 5.8–8.0. In addition, 0.8% d-glucose, 1 mg/mL glucose oxidase (266,600 units/g, Sigma), 0.4 mg/mL catalase (2000–5000 units/mg, Sigma), and 4 mM Trolox were added to the folding buffer for preventing the photobleaching.

DNA Constructs

All the DNA oligonucleotides were purchased from Sangon Biotech (Shanghai, China). The detailed sequences and modified positions are listed in Table S1. For the DNA constructs used in single-molecule measurements, 5 nM DNA was annealed by incubating a 1:2 mixture of the stem and ssDNA strands at 95 °C for 10 min and then slowly cooled down to 16 °C in approximately 7 h. Excessive use of biotin-free strands reduces the possibility of nonannealed strands anchoring to the coverslip surface.

FRET Melting

FRET-melting experiments were conducted with the FAM-TAMRA dual-labeled oligomers listed in Table S1 using a Rotor-Gene Q real-time PCR machine (QIAGEN). The oligonucleotides were generally measured at a 0.5 μM strand concentration in different kinds of buffers unless otherwise specified. The emission of the FAM fluorophore was normalized between 0 and 1, and the melting temperature Tm was determined as the temperature at which the normalized emission equaled 0.5.

Single-Molecule Fluorescence Data Acquisition and Data Analysis

The single-molecule fluorescence experiment was performed as described previously.[23] Streptavidin (10 μg/mL) was added to the microfluidic chamber and incubated for 10 min. Then, 50 pM DNA was added to the chamber where it was immobilized for 10 min. The free DNA was removed by washing with the folding buffer. Afterward, the chamber was filled with the folding buffer containing an oxygen-scavenging system (0.8% d-glucose, 1 mg/mL glucose oxidase, 0.4 mg/mL catalase, and 4 mM Trolox). We used an exposure time of 100 ms for all single-molecule measurements at a constant temperature of 22 °C. The FRET efficiency was calculated using IA/(ID + IA), where ID and IA represent the donor and acceptor intensities, respectively. Basic data analysis was carried out by scripts written in MATLAB. All fluorescent spots in each movie were selected unless the trace showed a poor signal: noise ratio or the intensity changes of the donor and acceptor did not match well. Each FRET histogram was constructed from more than 300 traces.

CD Spectropolarimetry

CD experiments were performed with a Bio-Logic MOS450/AF-CD optical system (Bio-Logic Science Instruments, France). A 3 μM solution of DNA sequences listed in Table S1 was incubated at 95 °C for 5 min and then slowly cooled down to room temperature in about 7 h. CD spectra were recorded in the UV (220–320 nm) regions in 0.75 nm increments with an averaging time of 2 s at 25 °C. The melting temperatures of i-motif structures can be obtained by heating the sample into 95 °C for 5 min and then slowly cooling down to 25 °C with 2 °C per step. The duration of each step was 2 min, and the CD spectrum was recorded at each constant temperature within about 50 s. The changes in the peak value of the CD spectrum versus temperature were then plotted. The melting temperature Tm was determined as the temperature at which the normalized peak value equaled 0.5.

Native Electrophoresis

The nondenaturing polyacrylamide gel electrophoresis (PAGE) was performed to distinguish the formation of higher-order DNA structures. The same amount of DNA was loaded into each well in the 15% gels which were run for 1 h at 120 V in the 1× tris-acetate-EDTA buffer (pH 7.5). The electrophoresis tank was placed on ice to ensure the low temperature in the gel system. The fluorophore-labeled DNA was then directly observed with a ChemiDoc MP imaging system using the FAM channel (Bio-Rad).

Results

Potassium Ion Displays Opposite Effects on the Thermal Stability of i-Motif DNA in Different Buffer Systems

The formation of i-motif requires the hemi-protonated cytosine–cytosine pairs CH+: C. Monovalent cations can bind to nucleic acid molecules and affect their physical properties.[24] However, different labs have reached different conclusions about how the ubiquitous presence of monovalent cations impacts the folding of i-motif DNA. Here, we used potassium ions with a concentration gradient to systematically dissect the effect of monovalent cations in different buffer systems. The C-rich sequences from the Bcl2 oncogene promoter, human telomere, and ILPR promoter (referred to as Bcl2-IM, hTel-IM, and ILPR-IM) were selected as they originate from different genomic regions and are linked to different biological/disease functionalities. Besides, these sequences were often used in previous i-motif studies.[25]

First, the thermal stabilities of i-motif structures were characterized by the FRET-melting assay. A donor fluorophore (FAM) and an acceptor fluorophore (TAMRA) were labeled at the ends of i-motif DNA, as shown in Figure A. The unfolding of the i-motif structures by heating will lead to increases in donor intensity. Figure B shows the FAM intensities of Bcl2-IM measured in 50 mM MES, pH 5.8 with 0–500 mM KCl. Then, the emission of FAM was normalized between 0 and 1, as shown in Figure C, and the value at which the normalized emission equals 0.5 was determined as the melting temperature Tm. With the increase in K+ concentrations from 0 to 500 mM in 50 mM MES, pH 5.8, the Tm value of Bcl2-IM decreases continuously from 62.3 to 57.3 °C. In contrast, the Tm value increases from 44.0 to 52.9 °C when the K+ increases from 0 to 500 mM in 50 mM PB, pH 5.8 (Figure D,E). Figure S1 further demonstrates that Tm values did not change with the variation of DNA concentrations from 0.25 to 2 μM. Therefore, the unimolecular i-motif structures should be predominantly formed in our experimental condition.

Figure 1

Thermal stability of Bcl2 i-motif DNA in different buffer systems at pH 5.8 and pH 6.6. (A) Schematic design of the FRET-melting assays. The FAM intensity is high with the ssDNA state of i-motif, while the folding of i-motif DNA generates a low FAM emission due to the energy transfer to TAMRA. (B,C) Original emission and normalized FAM intensity of Bcl2-IM at 25–95 °C in 50 mM MES (pH 5.8) with 0–500 mM KCl. The temperature at which the normalized emission equals 0.5 was defined as Tm. (D,E) Original emission and normalized FAM intensity of Bcl2-IM in 50 mM PB (pH 5.8). (F,G) Tm values of Bcl2-IM with the increase of the K+ concentration in different buffer systems at pH 5.8 and pH 6.6.

The CD spectra in Figure S2 confirmed the formation of i-motif structures at room temperature in both the MES and PB buffers at pH 5.8. Besides, the melting temperature Tm can also be obtained by the CD-melting assay from 25 to 95 °C. The Tm values at 0 mM K+ were ∼59.0 °C in MES and ∼40 °C in PB (Figure S2A–C), a bit lower than those determined from the FRET-melting assay. Although there is a small discrepancy between these two methods, the Tm values show clear decreases when the K+ concentration increases in MES (Figure S2D–F), which is consistent with the FRET-melting, and confirmed our results in Figure F. We also characterized the influences of Na+ on the thermal stability of i-motif at pH 5.8 (Figure S3). Interestingly, the Tm first decreases, then increases, and finally maintains at a relatively constant level in MES once Na+ is above 100 mM. This phenomenon is consistent with the previous report.[22] However, in the PB system, the Tm continues to increase with the increase in Na+ concentrations, which is similar to that in K+ (Figure F). Once again, the different effects of monovalent ions on i-motif thermal stability in different buffer systems were observed.

As K+ demonstrate totally different effects on the thermal stability of i-motif in the above-mentioned two buffer systems, we then continued to select three other kinds of buffers including Bis-Tris, SSC, and SCB to further examine the influence of K+ on Bcl2-IM folding. Due to the different buffer ranges (MES, pH 5.5–6.7; PB, pH 5.5–8.5; Bis-Tris, pH 5.8–7.2; SSC, pH 3.0–6.6; and SCB, pH 5.0–7.6), we chose the two pH values in that they all have buffering effects, pH 5.8 and pH 6.6, for the test. We observed that in different buffer systems even at the same pH value and ion concentration, there are significant differences in the Tm values (Figure F,G), suggesting that the solute molecules may also interact with i-motif DNA. More importantly, Figure F confirmed that K+ in different buffers indeed has opposite effects on the thermal stability of Bcl2-IM. Specifically, at pH 5.8, in MES and Bis-Tris, Tm values decrease continuously when the K+ concentration increases; however, in PB, SSC, and SCB, Tm values increase. At pH 6.6, Tm continuously decreases only in Bis-Tris. In other buffers, the values of Tm either continuously rise (PB, SSC, and SCB) or first decrease slightly and then increase (MES). Similar phenomena were also found in the ILPR promoter and human telomeric i-motif sequences, as shown in Figure S4.

Altogether, the above-mentioned pieces of evidence indicate that high concentrations of potassium ions in Bis-Tris always have a destructive effect on i-motif structures, while the results are opposite in PB, SSC, and SCB. As for MES, depending on the pH and sequences, potassium ions may have different effects.

High Concentrations of Potassium Ions in PB Promote the Folding of i-Motif DNA

A number of studies have established that the monovalent alkaline metal ions such as Na+ and K+ have a destructive effect on the i-motif structures at the low pH.[26,27] Given that we detected the opposite results in the above-mentioned section, here, we chose the typical PB system in which K+ increases the thermal stability of i-motif structures and employed smFRET to further monitor the conformational changes of i-motif structures at a series of pH and K+ concentrations. As shown in Figure A (Bcl2-IM) and 2E (ILPR-IM), a donor Cy3 and a receptor Cy5 were labeled at the 3′ end of i-motif DNA and the sixth nucleotide inside the stem duplex, respectively. Therefore, the changes in FRET values can sensitively reflect the conformation changes in the i-motif structures. In addition, there is a 2-nt linker between the duplex and i-motif DNA to avoid the interactions between them according to the previous studies.[28,29] Particularly, the formation of G4 and i-motif conformations was reported to destabilize the proximal duplex DNA when there was little space between them.[30] A pocket may also be formed in the quadruplex–duplex interface when they were proximal to each other.[31]

Figure 2

High concentration of potassium ions promotes the i-motif DNA folding in PB. (A,E) Schematic diagrams of Bcl2-IM and ILPR-IM DNA labeled with Cy3 and Cy5. Changes in DNA conformation will cause changes in the distance between the donor and acceptor fluorophores, leading to the variation in FRET efficiency. (B,F) FRET distributions at pH 5.8. The FRET peaks in varying concentrations of potassium were maintained at ∼E0.9 (Bcl2-IM) or ∼E0.8 (ILPR-IM). (C,G) FRET distributions at pH 7.0. Multiple peaks were detected, and the FRET distributions significantly shift to the right with the increase in the K+ concentration. (D,H) FRET distributions at pH 8.0. Two peaks were detected in Bcl2-IM at ∼E0.4 and ∼E0.6, and the FRET distributions shift to the E0.6 peak with the increase in the K+ concentration. A single peak was detected in ILPR-IM at ∼E0.4.

To observe how K+ affects the i-motif structures at different pHs, we used 50 mM PB with 50–750 mM KCl under three representative conditions at the acidic pH 5.8, neutral pH 7.0, and alkaline pH 8.0, respectively. At pH 5.8, only one dominant population at ∼E0.9 or ∼E0.8 is presented in the FRET histograms (Figure B,F), reflecting the folding of i-motif structures. Since acidic conditions are suitable for the formation of i-motif structures, further increase in the K+ concentration only slightly made the distributions narrower, that is, promoting the homogeneous i-motif folding. At pH 7.0, both Bcl2-IM and ILPR-IM show multiple folding states at 50 mM K+, reflected by the multiple peaks in the FRET distributions (Figure C,G), which is consistent with the observation that i-motif structures coexist with the partially folded structures.[32] More importantly, the FRET distributions shifted to the higher bands with the increase in K+ from 50 to 750 mM. Figure S5 shows that the FRET distributions only slightly shift to the right in the poly-T ssDNA with the increase in the K+ concentration, possibly owing to the electrostatic shielding effect of K+ on the negatively charged DNA backbones. Therefore, the significant FRET changes in Figure C,G should reflect the formation of higher-order DNA structures. At pH 8.0, the i-motif sequences mainly exist as partially folded states at high K+ concentrations, reflected by the increased FRET values (Figure D,H). Altogether, the above-mentioned results indicate that high concentrations of potassium ions can indeed promote i-motif DNA to form higher-order structures in PB, consistent with our observations in Figure .

Increase in Thermal Stability Does Not Represent the Formation of the i-Motif Structure

One may query whether the increase in Tm values and change in FRET distributions mentioned above imply that K+ in the PB can promote the ssDNA or partially folded states to fold into the i-motif structure. To answer this question, we used CD spectroscopy to detect the folding status of Bcl2-IM, hTel-IM, and ILPR-IM, which were annealed in the PB solution at different pH values and K+ concentrations.

Figure A–C demonstrates that different from the random ssDNA (Figure D), Bcl2-IM, hTel-IM, and ILPR-IM show the typical i-motif CD spectrum at pH 5.8 with 100–750 mM K+, which is indicated by positive ellipticity at ∼285 nm and negative ellipticity at ∼255 nm.[33] However, in the neutral and alkaline environments, the CD spectra of those sequences may reflect some partially folded structures rather than ssDNA or the i-motif structure. With the increase in the potassium concentration, we did not observe the shift of CD spectra to the typical i-motif band. This illustrated that although the high concentration of K+ improved the folding of i-motif DNA sequences at neutral and alkaline pH, as shown in Figure , it did not induce the partially folded structures to transform into i-motif. It is likely that some unknown types of higher-order structures were formed in these i-motif sequences at neutral pH. Figure E further verifies the formation of higher-order structures in i-motif DNA using the native PAGE. At 0 mM K+, there is only one DNA band at both pH 7.0 and pH 8.0 in PB, reflecting that no additional structures were formed. However, with the increase in the K+ concentration, a second DNA band appeared, especially at or above 250 mM K+. Once again, this observation confirmed that a high K+ concentration can promote the formation of higher-order DNA structures in i-motif sequences.

Figure 3

Formation of higher-order DNA structures in i-motif sequences at neutral pH in PB. (A–C) CD spectra of Bcl2-IM, hTel-IM, and ILPR-IM at different pH and K+ concentrations. Only at pH 5.8, the typical CD spectrum of i-motif structures with the peak at ∼285 nm and valley at ∼255 nm can be observed. At pH 7.0 and pH 8.0, the CD spectrum indicates the formation of other types of higher-order DNA structures. (D) CD spectrum of the 20-nt ssDNA at pH 7.0 with 100 and 250 mM K+. No significant CD signal can be observed. (E) Native PAGE image of FAM and TAMRA dual-labeled Bcl2-IM DNA, which was annealed at pH 7.0 at varied K+ concentrations. The gel was visualized by the FAM fluorophore. With the increase in the K+ concentration from 0 to 750 mM, a second band appears, which reflects the formation of the higher-order DNA structures. As a control, Bcl2-IM only displays one band at pH 8.0 when there was no K+, similar to that at pH 7.0 with no K+.

As mentioned earlier, Ag+ ions could replace the protonation of H+ to induce the i-motif sequence to fold into the i-motif structure at room temperature and even physiological pH.[18] Obviously, K+ ions did not have this effect in PB. However, K+ can stabilize the i-motif formed at acid conditions and promote the C-rich sequences to fold into higher-order structures other than i-motif at neutral and alkaline conditions. In addition to address the conformation in PB, we also measured the CD spectrum of Bcl2-IM in MES and Bis-Tris. Figure S6A shows that, at acidic pH, Bcl2-IM predominantly folded into the i-motif structures (pH 7.0 cannot be obtained in MES due to the limit of the buffer range). Interestingly, in the Bis-Tris buffer (Figure S6B), the i-motif structures were formed at pH 5.8 regardless of the K+ concentration; however, at pH 6.6 and pH 7.0, the typical i-motif signature can only be observed at a low K+ concentration. At a high K+ concentration, the peak shifted significantly to the left, suggesting the formation of other types of structures. This is consistent with the negative effect of K+ on i-motif thermal stability in the Bis-Tris buffer described above.

C-Hairpins May Be Formed at High Potassium Concentrations

Previous studies have proposed that C-hairpins are likely to be the intermediate states in i-motif DNA folding.[33] Therefore, the higher-order structures formed at pH 7.0–8.0 in Figure A–C may be the C-hairpins by C+: C base pairs. To verify this speculation, we conducted the following experiments to examine the effect of potassium ions on the thermal stability of the i-motif-derived C-rich sequences (Figure A). The hTel-IM sequence was selected due to the uniform loop lengths.

Figure 4

Formation of higher-order DNA structures in the partial i-motif sequences in PB. (A) Sequences of mutant hTel-IM. Three to five cytosines in two or three C-tracts are included. (B–D) Tm values of the above-mentioned sequences at pH 5.8, pH 6.6, and pH 7.0 in different concentrations of K+ determined by FRET-melting. At pH 6.6 or 7.0 and low K+ concentrations, there are several sequences whose Tm values cannot be precisely determined due to the low thermal stabilities. (E) Native PAGE image of FAM and TAMRA dual-labeled C4L3 and 2C1 at varied K+ concentrations. The gel was visualized by the FAM fluorophore. In 500 mM K+, the DNA band migrated slower, reflecting the formation of the higher-order structures. Besides, the weakened FAM intensity further suggests the folding of these sequences at high K+ concentration as FAM may transfer its energy to the TAMRA.

First, we measured the thermal stability of six mutant hTel-IM C-rich DNA sequences in PB at pH 5.8, 6.6, and 7.0, respectively. The choice of these three pH values was based on the conditions used in measuring the Tm values of the original i-motif in Figures and S4. With the increase in the K+ concentration, the Tm values of all these six sequences increase at all three pHs (Figure B–D), reflecting the formation of C-hairpins. Figure E further confirms the formation of higher-order structures in the two selected C-rich DNA sequences at high potassium concentrations according to the slow migration in native PAGE. It is worth noting that the tendency of Tm change in the above-mentioned sequences is similar to the hTel-IM, further reflecting that high potassium concentrations can promote the folding of C-hairpins. In addition, we were surprised to find that the Tm values of these sequences were all higher than those of the original hTel-IM (Figure B). Therefore, we speculate that the C-rich sequences may fold into some intermolecular structures by C+: C base pairs besides the intramolecular C-hairpins in the solution.

Second, we picked three hTel-IM-derived sequences and evaluated their folding status by analyzing the FRET distributions at different pH and K+ concentrations in PB (Figure S7). At pH 5.8 which is conducive to C+: C base pairs, the FRET distribution is basically unchanged or slightly shifts to the right as the potassium concentration increases, indicating that the C-hairpins may have little change in the folding conformation. At pH 6.2 and pH 7.0, the FRET distributions shift obviously to the right with the increase of the K+ concentration, indicating that K+ can promote the folding of C-hairpin structures. This observation is consistent with the increase in thermal stability (Tm values) at high K+ concentrations, as shown in Figure .

Taken together, the above-mentioned pieces of evidence indicate that potassium ions in PB buffer systems can enhance the thermal stability of the higher-order structures including both the i-motif and C-hairpins that are held together by the C+: C base pairs. As the i-motif sequences cannot fold into i-motif structures at the neutral and slightly alkaline conditions (Figure ), the C-hairpins are most likely to form instead.

Influence of Complementary Strands on i-Motif DNA at Different pH and Ion Concentrations

Several previous studies have focused on the competence between i-motifs, G4 structures, and the duplex DNA. For instance, the 1:1 mixture of the G-rich and C-rich sequences at acidic pH produced predominantly the i-motif and G4 structures, respectively.[34] Besides, both the c-MYC G4 and i-motif were shown to present simultaneously in opposite stands with slight displacement with each other.[35] Later, optical tweezers demonstrated that G4 and i-motif are mutually exclusive in both the c-MYC and ILPR promoters governed by the steric hindrance.[36,37] Recently, both the two structures were reported to be formed when the sequences are offset in the two strands.[38] Therefore, we further characterized the thermal stability of i-motifs under the influence of complementary strands using the FRET-melting assay in PB at different ionic and pH conditions.

The fluorophore-labeled i-motif sequence and unlabeled G4 sequence were mixed with different ratios from 1:0 to 1:3, and the schematic experimental design is shown in Figure A. The FAM fluorophore is in low intensity once i-motif is well folded. If the G4 strands form the duplex with i-motif, the FAM intensity may go up initially due to the DNA linearization. During the heating process, the i-motif is partially or completely unfolded, leading to an increase in FAM intensity. Meanwhile, the unfolded i-motif strand will form the duplex with the G4 strand, and the linearization will lead to further increases in FAM intensity. When the temperature further goes up, the duplex denatures and the FAM intensity will decrease once the ssDNA strand is released.

Figure 5

At low pH, the free complementary G-rich strands display little impact on the folding of i-motif DNA. (A) Schematic diagram of the FRET-melting experimental design. (B–D) Changes in FAM intensity during the heating process in PB. (E–G) Normalized FAM emission in the presence of G4 strands. Only at high ion strength and high G4 DNA concentrations, the decreases in Tm values can be observed.

At pH 5.8, the FAM intensity did not change when G4 strands were added to the i-motifs, indicating that the folding of i-motifs was not influenced by the free G4s at the temperature lower than 35 or 40 °C (Figure B–D). Interestingly, there was a decrease in FAM intensity once the temperature went beyond ∼80 °C at 250 and 500 mM K+ in the presence of G4 strands, which should be due to the denaturation of the duplex DNA. The normalized emissions in Figure E–G indicate that the Tm of i-motif at 0 mM K+ barely changed with the addition of G4 strands. At 250–500 mM K+ (high K+ concentrations may promote the formation of duplex DNA), only high G4 concentrations can lead to the decreases in i-motif stability. We next examined the effects of G4s at pH 7.0 (Figure S8). To our surprise, with the addition of G4 strands, the FAM intensity significantly increased even at room temperature before the heating (Figure S8A–C). This phenomenon strongly reflects the unfolding of i-motif structures in the presence of the complementary strands (Figure S8D).

Altogether, the above-mentioned results indicate that, at acidic conditions, the free complementary strands show little effects on i-motif folding; however, at neutral conditions, the formation of duplex DNA is favored regardless of the K+ concentrations. It is worth noting that the G4 and i-motif strands in our experiments were in free states without the steric hindrance with each other;[36−38] therefore, the pure effects of pH and ions could be obtained.

Discussion

Relying on the research of Schildkraut and Lifson, we learned that alkaline metal ions at moderate concentrations can enhance the thermal stability of duplex DNA.[39] G4 DNA structures can also be stabilized by K+, Na+, or other ions in their folding topology and thermal stability.[40,41] All these pieces of evidence suggest that K+ may have positive effects on the thermal stability of higher-order DNA structures. Interestingly, Zhang et al. pointed out that Ag+ can stabilize the i-motif structures at neutral pH; however, at acidic pH, Ag+ decreases the thermal stability of i-motif structures. In their experiments, they selected MES to create an acidic environment and used MOPS to simulate the physiological pH. Sixteen kinds of metal ions such as Na+, K+, Li+, and Mg2+ were added to MES solutions at pH 6.5. As the ion concentration increased, they found that almost all metal ions have a destructive effect on the thermal stability of i-motif structures in MES,[26] consistent with our results in Figure B,C. Besides, we speculate that the opposite effects of Ag+ on the thermal stability of i-motif structures under neutral and acidic pH in their study may also be related to the influence of different solute molecules in different buffer systems, as observed in Figure .

The proposed mechanisms of the opposite effects of K+ on i-motif DNA folding are shown in Figure . It is worth noting that in most conditions, the Tm values of i-motif at low K+ concentrations in MES and Bis-Tris buffers are higher than those in other buffers at low pH (Figures and S4). These pieces of evidence suggest that these solute molecules may stabilize the folding of i-motif structures which are held together by the intercalated C+: C base pairs. When the monovalent cations have a negative effect on the thermal stability of i-motif structures, it is easy to conjecture that they may compete with hydrogen ions (Figure A). As we know, the most important factor that makes i-motif DNA fold into the specific topology and stabilizes i-motif structures is the semi-protonated hydrogen bonds C+: C.[42] The same positive charge carried by the monovalent cations may prohibit the protonation of the cytosine pairs. As a result, i-motif sequences may exist as the partial folded C-hairpins or ssDNA status at a high K+ concentration in the buffer systems such as MES and Bis-Tris.

Figure 6

Proposed functions of potassium ions on i-motif DNA folding in different buffer systems. (A) Hemi-protonated C+: C base pairs were destabilized by K+ in buffers such as MES and Bis-Tris, causing the complete or partial disintegration of the i-motif structure. (B) Under the acidic condition in buffers such as PB, K+ may neutralize the electrostatic repulsion generated by the i-motif. Under neutral and slightly alkaline conditions, i-motif DNA tends to exist as partially folded C-hairpins or interchain duplex by C+: C base pairs and potassium ions may promote the formation of these structures.

In addition to competing with hydrogen ions, the positively charged K+ can compensate for the electrostatic repulsion between the negatively charged nucleic acid structures to a certain extent, thus displaying a positive effect on the stability of the i-motif structures in buffers such as PB (Figure B). In fact, the most important factor that determines whether the i-motif structure can be formed is pH;[43] therefore, the i-motif structure can be readily folded under acidic conditions. The existence of potassium ions in PB may shield the negative charge carried by the nucleic acid backbone, further stabilizing the i-motif structures, as shown in Figures E and 2B. At the neutral and alkaline conditions, i-motif DNA mainly exists as intermediate states such as C-hairpins or intermolecular structures (Figure ). The increase of potassium ions may promote the formation of those structures (Figure ), in a similar way as the stabilization effects of cations on the duplex DNA by neutralizing the negative charges on DNA backbones.[44] Although we are not able to exactly tell the structure of the intermediate states at the current stage, in the future, the molecular simulations may be used to substantiate the formation of those structures in i-motif sequences according to the previous studies on G-triplex and G-hairpin in G4 DNA folding.[45,46]

Conclusions

In this report, we directly revealed the opposite effects of potassium ions on the thermal stability of i-motif DNA by systematically investigating the thermal stability and folding status of i-motif under different K+ concentrations in several buffer systems. In particular, K+ in PB promotes the folding of i-motif sequences into higher-order structures. Further studies revealed that higher thermal stability at the neutral and alkaline conditions was not equivalent to the formation of i-motif structures. Instead, other structures such as the C-hairpins or intermolecular duplex by C+: C base pairs may be formed. The opposite effects of monovalent cations we observed are the results of the combination of multiple factors such as C+: C base pairs, solvent molecules, and electrostatic interaction. Our results may broaden our understanding about the formation and thermal stability of i-motif DNA structures both in vitro and in vivo.