Effects of non-CpG site methylation on DNA thermal stability: a fluorescence study

Abstract

Cytosine methylation is a widespread epigenetic regulation mechanism. In healthy mature cells, methylation occurs at CpG dinucleotides within promoters, where it primarily silences gene expression by modifying the binding affinity of transcription factors to the promoters. Conversely, a recent study showed that in stem cells and cancer cell precursors, methylation also occurs at non-CpG pairs and involves introns and even gene bodies. The epigenetic role of such methylations and the molecular mechanisms by which they induce gene regulation remain elusive. The topology of both physiological and aberrant non-CpG methylation patterns still has to be detailed and could be revealed by using the differential stability of the duplexes formed between site-specific oligonucleotide probes and the corresponding methylated regions of genomic DNA. Here, we present a systematic study of the thermal stability of a DNA oligonucleotide sequence as a function of the number and position of non-CpG methylation sites. The melting temperatures were determined by monitoring the fluorescence of donor-acceptor dual-labelled oligonucleotides at various temperatures. An empirical model that estimates the methylation-induced variations in the standard values of hybridization entropy and enthalpy was developed.

Figure 1.

Sequences of the dual-labelled probe and the differentially methylated target oligonucleotides. The methylcytosines are indicated with a red ‘M’.

Figure 2.

Denaturation data obtained from the measurement of UV absorbance at 260 nm versus temperature, T, for samples (a), (b), (e) and (i) (see Figure 1 for details). The solid lines represent the best fit of the experimental data to Equation (2a).

Figure 3.

Steady-state fluorescence signal, F, measured as a function of temperature, T, for samples containing 50 nM concentrated free TAMRA (squares), TAMRA-labelled ss-DNA (circles) or TAMRA-labelled ds-DNA (stars, the probe was hybridized to target A), see Figure 1.

Figure 4.

Sigmoidal denaturation plot obtained from the fluorescence versus temperature, T, data of Figure 3 by dividing the ds-DNA sample data, F_DS, by the corresponding single-stranded probe data, F_SS, as detailed in the text. The solid line represents the best fit of the experimental data to Equation (2b).

Figure 5.

Fractional concentrations of ss-DNA, f_SS and ds-DNA, f_DS, obtained as a function of temperature, T, through the fitting of the time-resolved fluorescence decay data of a ds-DNA sample obtained by hybridization of the probe to target (a) of Figure 1, as detailed in the text. The solid lines represent the best fit of the experimental data to Equation (2c).

Figure 6.

Comparison of the melting temperature values obtained for centrally methylated samples (a), (b), (e) and (i) of Figure 1 (0–10 methylations) in 150 mM PBS buffer with the different procedures detailed in the text. Crosses: melting temperatures, T_den, (black) and total denaturation temperatures, T_td, (cyan) obtained from the sigmoidal patterns obtained by processing the UV absorption data (Figure 2). Squares: melting temperatures, T_steady-state, obtained by fitting the sigmoidal patterns obtained by processing the steady-state fluorescence data (Figure 4). Circles: melting temperatures, T_timing, provided by fitting the denaturation plots obtained from the time-resolved measurements (see Figure 5). Stars: temperatures, T_peak, at which the steady-state fluorescence intensity patterns measured for the ds-DNA samples display a peak (Figure 3). Inset: differences between T_peak and T_steady-state (squares) or T_timing (circles) at different methylation extents.

Figure 7.

Values of T_peak, measured as a function of the ionic strength, I, for the duplexes formed by the dual-labelled probe with the non-methylated target oligonucleotide (a) of Figure 1 (squares), and with the target oligonucleotides (b), (e) and (i) of the same figure, containing one (circles), five (stars) and ten (diamonds) centrally located methylcytosine nucleotides, respectively. Inset: semi-logarithmic plot of the same data, together with their linear fits.

Figure 8.

Dependence of T_peak on the position of methylation sites for oligonucleotides with a fixed number of methylation sites. The denaturation experiments were performed at different ionic strength, I, values using duplexes that contained the same number of methylcytosines, namely, 1 (panel (A), target oligonucleotides (b), (c) and (d) in Figure 1), 5 (panel (B), target oligonucleotides (e), (f), (g) and (h)) and 10 (panel (C), target oligonucleotides (i), (j) and (k)), either in the central position (squares), next to the oligonucleotide ends (dots and circles for the 3′ and 5′ ends, respectively), or evenly distributed throughout the oligonucleotide (stars). For 10 methylation sites, the effects of methylation in the lateral position were only evaluated for one oligonucleotide, with 5 methylcytosines at each end (see Figure 1 for details).

Figure 9.

Methylated DNA melting temperature predictions compared to the experimental results. The predicted melting temperature, T_calc, is plotted versus the experimentally measured melting temperature, T_exp = T_peak – 5°C, at six ionic strengths represented by various colours. The continuous line represents the x = y line to emphasize the correspondence between the predicted and measured values.