Effects of CpG methylation on the thermal stability of c-kit2, c-kit*, and c-kit1 G-quadruplex structures

Highlights

• •CpG methylation at C1 or C11 decreases stability of c-kit2 G4 in Na⁺ and Mg²⁺.
• •CpG methylation at C5 increases stability of c-kit2 G4 in Na⁺ and Mg²⁺.
• •Oligomerization of c-kit2 G4-forming oligonucleotide is induced by C1 methylation.

Abstract

In genomic DNA, G-quadruplex (G4)-forming DNA can form either a duplex or G4 structure, suggesting that understanding the factors regulating G4 formation is important for revealing the cellular functions controlled by G4 formation. Cytosine DNA methylation in the CpG islands is known to play an important role in transcriptional regulation. Additionally, CpG methylation increases the thermal stability of G4 structures such as BCL2 and VEGF G4. In this study, we evaluated the effects of CpG methylation in three G4 structures (c-kit2, c-kit*, and c-kit1) produced by the c-KIT promoter. Each was analyzed using circular dichroism (CD) melting analysis. The results demonstrate that CpG methylation does not alter the thermal stability of c-kit2 G4 structure when formed in the presence of K⁺; a single-CpG methylation at C1 or C11 decreases the thermal stability of any c-kit2 G4 structure formed in the presence of Na⁺ and Mg²⁺ while methylation at C5 increases the thermal stability; CpG methylation does not alter the thermal stability of c-kit1 or c-kit* G4 structures formed in the presence of K⁺; and the c-kit1 and c-kit* G4-forming oligonucleotides do not form G4 structures in the presence of Na⁺ and Mg²⁺. These results provide important clues for understanding the regulatory mechanisms underlying the formation of CpG methylation-induced G4 structures.

1. Introduction

Noncanonical DNA structures such as cruciform, triplex, i-motifs, and G-quadruplexes (G4) exist in genomic DNA [1]. The G4 structure, formed by the stacking of two or more planar guanine tetrads, is the most analyzed noncanonical structure [2]. G4-forming sequences have been identified using G4 ligands coupled with high-throughput sequencing or DNA microarray technologies in genomic DNA [3], [4], [5], [6], [7] and transcribed RNA [8,9]. G4 structure formation is involved in various cellular functions, such as transcription [10], [11], [12], translation [13], splicing [14], replication [15], and telomere maintenance [16]. In genomic DNA, G4-forming DNA can form either a duplex or G4 structure, indicating that understanding the factors regulating G4 formation is important for elucidating the regulatory mechanisms and cellular functions controlled by G4 formation.

G4 structures are stabilized by monovalent and divalent cations, such as K⁺, Na⁺, and Mg²⁺ [17]. K⁺ concentrations in the cell alter the transcription of genes containing G-rich regions [18]. G4 structures are stabilized by G4-binding proteins. Nucleolin, a multifunctional phosphoprotein, facilitates c-myc G4 structure formation [19]. Luciferase reporter assays have demonstrated that nucleolin overexpression reduces c-myc promoter activity. In contrast, the c-myc G4 structure is dissolved by NM23-H2, which induces c-myc promoter activation [20].

Other factors that alter G4 formation are epigenetic DNA modifications, such as 5-methylcytosine at CpG dinucleotides and N⁶-methyladenine. CpG methylation stabilizes the d(CGCG₃GCG) oligonucleotides [21], FMR1 repeats [22], C9orf72 repeats [23], BCL2 [24], and VEGF [25,26] G4 structures, whereas it destabilizes the MEST G4 structure [27]. The binding activities of G4 structures to G4-binding proteins are regulated by CpG methylations, suggesting that CpG methylations on G4 structures may be transcriptional regulations [28]. Three G4-forming sequences (c-kit2, c-kit*, and c-kit1) have been identified in c-KIT promoters [29], [30], [31], [32]. Recently, we reported that N⁶-methyladenine modification destabilizes c-kit1 G4 structures that contain two A–G base pairs in a five-residue stem loop [33]. N⁶-methyladenine modification destabilizes Watson–Crick base pairing [34,35], suggesting that the destabilization of the five-residue stem loop by N⁶-methyladenine modifications reduces c-kit1 G4 structure stability.

In this study, we analyzed the effects of CpG methylation on c-kit2, c-kit*, and c-kit1 G4 structures. The structure of the wild-type c-kit2 G4-forming oligonucleotide (5′–CGGGCGGGCGCGAGGGAGGGG–3′) in solution has not been previously reported because it folds into multiple G4 structures in the presence of K⁺ [30]. Conversely, the structures of the c-kit2 G4 G21T (G21 is replaced by T) and G12T/G21T (both G12 and G21 are replaced by T) mutants in solution have both been reported [36]. In the presence of 20 mM KCl, the c-kit2 G4 G12T/G21T mutant folds into a monomeric parallel G4 structure, whereas the c-kit2 G4 G21T mutant folds into a dimeric parallel G4 structure in the presence of 100 mM KCl. The c-kit* G4-forming oligonucleotide (5′– GGCGAGGAGGGGCGTGGCCGGC–3′) folds into a chair-type antiparallel G4 structure with two G-quartets in the presence of K⁺ [32]. The c-kit1 G4-forming oligonucleotide (5′–AGGGAGGGCGCTGGGAGGAGGG–3′) folds into a parallel G4 structure with a five-residue stem loop in the presence of K⁺ [37]. Here, we investigated the direct effects of CpG methylation on c-kit2, c-kit*, and c-kit1 G4 structures.

2. Material and methods

2.1. Circular dichroism (CD) measurement

All high-performance liquid chromatography-purified oligonucleotides were purchased from the Tsukuba Oligo Service. The sequences and methylated CpG sites of c-kit2, c-kit*, and c-kit1 G4-forming oligonucleotides are shown in Table S1. Before use, 20 μM oligonucleotides were prepared in 50 mM Tris-HCl (pH 7.4). KCl, NaCl, or NaCl with MgCl₂ were added at the indicated concentrations. The oligonucleotides were denatured at 95 °C for 3 min in a Veriti Thermal Cycler (Thermo Fisher Scientific) and cooled to 25 °C for 30 min. The CD spectra of the oligonucleotides were measured using a J1500 CD spectrometer (JASCO) with a quartz cell (optical path length of 1 mm) at a scanning speed of 200 nm/min, at 220–320 nm and 25–95 °C (at 1 °C intervals). To determine the apparent melting temperature (T_1/2), molar ellipticities at a positive peak at 25 and 95 °C were set as 100% and 0%, respectively, and the curves were fitted using GraphPad Prism 7 software (GraphPad Inc.). The T_1/2 values were recorded as the temperature at which the normalized molar ellipticity was 50%.

2.2. Native polyacrylamide gel electrophoresis (PAGE)

Before use, 20 µM c-kit2 G4-forming oligonucleotides were heat-treated as described above in 50 mM Tris-HCl (pH 7.4), 20 mM NaCl, and 5 mM MgCl₂. The oligonucleotides (10 μL) were then electrophoresed on 15% polyacrylamide gels containing 20 mM NaCl and 5 mM MgCl₂ in running buffer (89 mM Tris, 89 mM borate, 20 mM NaCl, and 5 mM MgCl₂). For oligonucleotide markers, 5.4 μg 100-, 60-, 40-, and 20-mer polyT oligonucleotides were electrophoresed (Eurofins or Integrated DNA Technologies). The gel was stained with ethidium bromide and the oligonucleotides were visualized using WSE-5300 Printgraph CMOS I (ATTO).

3. Results

3.1. CpG methylation does not alter the thermal stability of c-kit2 G4 structures formed in the presence of K⁺

The thermal stability of G4 structures depends on the available monovalent and divalent cations during formation [17]. G4 structure topologies are classified into parallel, antiparallel, and hybrid structures that exhibit characteristic CD spectra. Typically, parallel-type G4 structures have positive and negative peaks around 264 and 245 nm, whereas antiparallel-type G4 structures have positive and negative peaks around 295 and 260 nm, respectively [2,38,39]. Previous reports have demonstrated that the CD spectrum of c-kit2 G4-forming oligonucleotides have strong and weak positive peaks at 263 and 295 nm, respectively, in the presence of 100 mM KCl, indicating that it folds into multiple G4 structures [30]. Therefore, the c-kit2 G4 structure in the presence of K⁺ has not been resolved whereas the mutated c-kit2 G4-forming oligonucleotide structure has been reported [36]. The G12T/G21T mutant folds into a monomeric parallel G4 structure in the presence of 20 mM KCl, and the c-kit2 G4 G21T mutant into a dimeric parallel G4 structure in the presence of 100 mM KCl.

To analyze the effects of CpG methylation on the thermal stability of wild-type c-kit2 G4 structures formed in the presence of K⁺, the CD spectra of the unmethylated and fully methylated wild-type c-kit2 G4-forming oligonucleotides were measured in the presence of 20 or 100 mM KCl. The spectrum of the oligonucleotides had positive peaks at 263 or 264 nm at 25 °C in the presence of 100 or 20 mM KCl, respectively (Fig. S1, S2). This indicates that CpG methylation does not alter the topology of G4 structures that were formed in the presence of 20 or 100 mM KCl. To analyze their thermal stability, CD spectra were measured from 25 to 95 °C at 1 °C intervals, and the CD melting analysis was performed to calculate the apparent melting temperature (T_1/2) using the molar ellipticities at 263 or 264 nm. In the presence of 100 mM KCl, the T_1/2 values of the unmethylated and fully methylated c-kit2 G4 structures were 71.4±0.6 and 70.6±2.2 °C (Fig. S3A), whereas those in the presence of 20 mM KCl were 60.6±1.8 and 59.2±0.5 °C, respectively (Fig. S3B). The CD melting curves of c-kit2 G4-forming oligonucleotides methylated once at either C1, C5, C9, or C11 were also assayed in the presence of 20 mM KCl. The CD melting curves and the T_1/2 values were not affected by single methylation, indicating that CpG methylation does not alter the thermal stability of G4 structures formed in the presence of K⁺ (Table 1, Fig. S2, Fig. S3B).

Table 1.

T_1/2 values of the c-kit2 G4 structures in the presence of 20 mM KCl or 20 mM NaCl with 5 mM MgCl₂.

CpG methylation site	T1/2 (°C)
	20 mM KCl	20 mM NaCl, 5 mM MgCl2
None (unmethylated)	60.6±1.8	51.5±0.9
C1	61.5±0.4	45.3±1.6*
C5	59.5±0.3	54.9±0.1*
C9	60.1±0.7	48.9±0.4
C11	59.6±0.1	47.6±0.3*
C1, C5, C9, C11 (fully methylated)	59.2±0.5	44.2±0.7*

N = 3; mean ± standard deviation (SD); * P < 0.005 vs. unmethylated G4.

3.2. CpG methylations change the thermal stability of c-kit2 G4 structures formed in the presence of Na⁺ and Mg²⁺

Na⁺ is a monovalent ion known to induce G4 formation. Given this, we measured the CD spectra of unmethylated and fully methylated c-kit2 G4-forming oligonucleotides in the presence of 20 mM NaCl. The CD spectra demonstrated that the oligonucleotides did not fold into the G4 structure (Fig. S4). Next, we measured the CD spectra of the oligonucleotides in the presence of 20 mM NaCl with 5 mM MgCl₂ because Mg²⁺ stabilizes parallel VEGF G4 structures formed in the presence of K⁺ [40]. This induced parallel VEGF G4 structures in the presence of Na⁺ [26]. The CD spectrum of the unmethylated c-kit2 G4-forming oligonucleotide had a strong positive peak at 260 nm at 25 °C in the presence of 20 mM NaCl with 5 mM MgCl₂ (Fig. 1A). This indicates that Mg²⁺ induces the folding of c-kit2 G4-forming oligonucleotides to parallel-type G4 structures in the presence of Na⁺. In contrast, the CD spectrum of the fully methylated c-kit2 G4-forming oligonucleotide had a strong positive peak at 262 nm at 25 °C in the presence of 20 mM NaCl with 5 mM MgCl₂ (Fig. 2A). This indicates that fully methylated c-kit2 G4-forming oligonucleotide folds into a parallel-type G4 structure; however, the folding state is different from that of the unmethylated oligonucleotide.

Fig. 1.

CD spectra for c-kit2 G4-forming oligonucleotides prepared in 50 mM Tris-HCl (pH 7.4) supplemented with 20 mM NaCl and 5 mM MgCl₂ from 25 to 95 °C. A, unmethylated; B, fully methylated; C, C1 methylated; D, C5 methylated; E, C9; methylated; F, C11 methylated c-kit2 G4-forming oligonucleotides.

Fig. 2.

CD melting analysis for methylated c-kit2 G4-forming oligonucleotides prepared in 50 mM Tris-HCl (pH 7.4) supplemented with 20 mM NaCl and 5 mM MgCl₂. All experiments were performed in triplicate and data are presented as the mean ± SD.

To analyze the thermal stability of the unmethylated and fully methylated c-kit2 G4 structures formed in the presence of 20 mM NaCl with 5 mM MgCl₂, CD melting analysis was performed using molar ellipticities at 260 nm. The melting curve of unmethylated c-kit2 G4 structures exhibited a single melting transition (Fig. 2). In contrast, the melting curve of the fully methylated c-kit2 G4 structure exhibited two melting transitions. The individual melting curves are shown in Fig. S5. These results indicate that fully methylated c-kit2 G4-forming oligonucleotides fold into at least two structures and that thermal unfolding is not a two-state pathway. Therefore, T_1/2, the temperature at which the normalized molar ellipticity was 50%, was calculated. The T_1/2 values for unmethylated and fully methylated c-kit2 G4 structures were 51.5±0.9 and 44.2±0.7 °C, respectively (Table 1). The destabilization effects of full methylation in the presence of 20 mM NaCl with 2 mM MgCl₂ were also confirmed (Fig. S6, S7; T_1/2 = 48.4±1.5 and 42.6±1.1 °C for unmethylated and fully methylated c-kit2 G4 structures, respectively). These results indicate that full methylation decreases the thermal stability of c-kit2 G4 structures formed in the presence of Na⁺ and Mg²⁺.

To explain the effects of each CpG methylation on the thermal stability of c-kit2 G4 structures, CD melting for c-kit2 G4-forming oligonucleotides single-methylated at C1, C5, C9, or C11 were performed. The CD spectra of the C1, C9, and C11 methylated c-kit2 G4-forming oligonucleotides had a strong positive peak at 260 nm at 25 °C in the presence of 20 mM NaCl with 5 mM MgCl₂ (Fig. 1C, E, F). In contrast, the CD spectra of the C5 methylated c-kit2 G4-forming oligonucleotide had a strong positive peak at 258 nm at 25 °C in the presence of 20 mM NaCl with 5 mM MgCl₂ (Fig. 1D). This indicates that the C5 methylated c-kit2 G4-forming oligonucleotide folding state is different from the unmethylated oligonucleotide folding state. The melting curve of the C5, C9, and C11 methylated c-kit2 G4 structures exhibited a single melting transition whereas the melting curve of the C1 methylated c-kit2 G4 structure exhibited two melting transitions (Fig. 2, Fig. S5). The T_1/2 values of the C1, C5, C9, and C11 methylated c-kit2 G4 structures were 45.3±1.6, 54.9±0.1, 48.9±0.4, and 47.6±0.3 °C, respectively (Table 1). We also confirmed that the inflection point temperatures of the melting curves were consistent with the T_1/2 values (Table S2). These results indicate that C1 and C11 methylation decreases the thermal stability of c-kit2 G4 structures, whereas C5 methylation increases the thermal stability.

CD melting analysis indicated that full or only C1 methylation induced c-kit2 G4-forming oligonucleotides to fold into at least two structures. To analyze the structures, native PAGE analyses was performed using polyacrylamide gels and a running buffer containing 20 mM NaCl and 5 mM MgCl₂. Although the main band was detected in all c-kit2 G4-forming oligonucleotides around 20-mer, an additional band corresponding to a 40-mer for the C1 and fully methylated c-kit2 G4-forming oligonucleotides was observed (Fig. 3). These results indicate that the C1 and fully methylated c-kit2 G4-forming oligonucleotides fold into both intramolecular G4 and oligomeric structures, consistent with the CD melting curves, which exhibit two melting transitions (Fig. 2).

Fig. 3.

PAGE analysis of methylated c-kit2 G4-forming oligonucleotides in the presence of 20 mM NaCl and 5 mM MgCl₂. Lane 1, unmethylated c-kit2 G4-forming oligonucleotides; lane 2, C1 methylated c-kit2 G4-forming oligonucleotides; lane 3, C5 methylated c-kit2 G4-forming oligonucleotides; lane 4, C9 methylated c-kit2 G4-forming oligonucleotides; lane 5, C11 methylated c-kit2 G4-forming oligonucleotides; lane 6, fully methylated c-kit2 G4-forming oligonucleotides; lane 7, 100-mer polyT; lane 8, 60-mer polyT; lane 9, 40-mer polyT; lane 10, 20-mer polyT.

3.3. CpG methylation does not alter the thermal stability of c-kit1 and c-kit* G4 structures formed in the presence of K⁺

The c-kit* G4-forming oligonucleotide, which contains three CpG sites, folds into a chair-type antiparallel G4 structure in the presence of K⁺ [32]. The CD spectrum of the c-kit* G4-forming oligonucleotide had a positive peak at 290 nm, which is characteristic of the antiparallel G4 structure in the presence of 100 mM KCl (Fig. S8). The CpG methylated c-kit* G4-forming oligonucleotides also exhibited CD spectral features with an antiparallel G4 structure which indicated that CpG methylation did not affect G4 structure topology (Fig. S8). CD melting analysis indicated that CpG methylation did not affect the thermal stability of the c-kit* G4 structures (Fig. S9, Table S3). In the presence of 20 mM NaCl and 5 mM MgCl₂, the c-kit* G4-forming oligonucleotides did not exhibit the characteristic CD spectrum of G4 structure indicating that the c-kit* G4-forming oligonucleotides did not form G4 structures in the presence of Na⁺ and Mg²⁺ (Fig. S10).

The c-kit1 G4-forming oligonucleotide, which contains one CpG site, folds into a parallel G4 structure with a five-residue stem loop in the presence of K⁺ [37]. The CD spectrum of the unmethylated and methylated c-kit1 G4-forming oligonucleotides had a positive peak at 263 nm, which is characteristic of the parallel G4 structure in the presence of 20 mM KCl. This indicates that CpG methylation did not affect G4 structure topology (Fig. S11). CD melting analysis indicated that CpG methylation did not affect the thermal stability of c-kit1 G4 structures (Fig. S12, Table S4). In the presence of 20 mM NaCl and 5 mM MgCl₂, the G4-forming oligonucleotides did not exhibit the characteristic CD spectrum of G4 structure indicating that the c-kit1 G4-forming oligonucleotides did not form G4 structures in the presence of Na⁺ and Mg²⁺ (Fig. S13).

4. Discussion

In this study, we demonstrated that CpG methylation affects the thermal stability of the c-kit2 G4 structure formed in the presence of Na⁺ and Mg²⁺, but does not affect that formed in the presence of K⁺. This indicates that the c-kit2 G4 structure formed in the presence of Na⁺ and Mg²⁺ is different from that formed in the presence of K⁺. In normal cells, K⁺ concentration is higher than that of Na⁺ suggesting that c-kit2 G4-forming oligonucleotide mainly folds into a K⁺-induced G4 structure. In contrast, the intracellular K⁺ concentration in cancer cells is lower than that in normal cells because K⁺ channels are overexpressed in cancer cells [18]. Moreover, the decreased intracellular K⁺ concentration affected G4 formation in the cells. These results suggest that CpG methylation affects c-kit2 G4 formation in cancer cells.

The c-KIT proto-oncogene encodes a tyrosine kinase receptor whereby gain-of-function mutations promote tumor formation [41]. The c-KIT promoter contains three G4-forming sequences that contain a transcription factor SP1 binding site [31], and c-kit2 G4 interacts with c-kit* G4 in an oligonucleotide containing both c-kit2 and c-kit* G4-forming sequences [42]. G4 ligands that stabilize c-kit G4 structures reduce c-KIT expression suggesting that G4 formation suppresses gene expression [43]. In this study, we demonstrated that the thermal stability of c-kit2 G4 structures formed in the presence of Na⁺ and Mg²⁺ changes depending on CpG methylation site. These results contribute to elucidating the methylated G4 transcriptional regulation mechanisms.

5. Conclusions

In this study, we investigated the effects of CpG methylation on the thermal stability of c-kit2, c-kit*, and c-kit1 G4 structures. In the presence of K⁺, CpG methylation did not affect the thermal stability and topology of c-kit2, c-kit*, and c-kit1 G4 structures. In the presence of Na⁺ and Mg²⁺, the c-kit* and c-kit1 G4-forming oligonucleotides do not form G4 structure whereas those of c-kit2 form parallel-type G4 structures. Complete C1, C5, C9, and C11 methylation decreased the thermal stability of c-kit2 G4 structures formed in the presence of Na⁺ and Mg²⁺. C1 or C11 methylation decreases thermal stability whereas C5 methylation increases thermal stability. Moreover, C1 methylation induces oligomer formation by c-kit2 G4-forming oligonucleotides. These results provide important clues for elucidating the full CpG methylation-induced G4 formation regulation mechanisms.

CRediT authorship contribution statement

Saowalak Laddachote: Methodology, Formal analysis, Investigation, Data curation, Writing – review & editing, Visualization. Rika Ishii: Methodology, Formal analysis, Investigation, Data curation, Writing – review & editing, Visualization. Wataru Yoshida: Conceptualization, Formal analysis, Data curation, Writing – original draft, Visualization, Supervision, Funding acquisition.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.