Detection and quantification of methylation in DNA using solid-state nanopores

Abstract

Epigenetic modifications in eukaryotic genomes occur primarily in the form of 5-methylcytosine (5 mC). These modifications are heavily involved in transcriptional repression, gene regulation, development and the progression of diseases including cancer. We report a new single-molecule assay for the detection of DNA methylation using solid-state nanopores. Methylation is detected by selectively labeling methylation sites with MBD1 (MBD-1x) proteins, the complex inducing a 3 fold increase in ionic blockage current relative to unmethylated DNA. Furthermore, the discrimination of methylated and unmethylated DNA is demonstrated in the presence of only a single bound protein, thereby giving a resolution of a single methylated CpG dinucleotide. The extent of methylation of a target molecule could also be coarsely quantified using this novel approach. This nanopore-based methylation sensitive assay circumvents the need for bisulfite conversion, fluorescent labeling, and PCR and could therefore prove very useful in studying the role of epigenetics in human disease.

Figure 1. Detection of methylated and unmethylated DNA using a solid-state nanopore.

(a) Schematic diagram of a nanopore setup. A focused electron beam of TEM is used to sculpt a nanopore in a thin (~20 nm) silicon nitride membrane; the nanopore chip is sealed between two fluidic cell chambers containing conductive electrolyte; a voltage is applied across this setup to induce the translocation of single dsDNA molecules through the pore as shown. (Inset) TEM image of typical ~4.2 nm diameter nanopore used in DNA measurements (scale bar is 10 nm). (b) Characteristic ionic current traces obtained from the translocation of mDLX1 (827 bp dsDNA with 36 potential CpG methylation sites). Traces were recorded in 600 mM KCl at pH 8.0 electrolyte at various voltage levels. (c) A typical DNA induced current blockade. Parameters of interest are the open pore current, I_O, residual blocking current, I_B, (occurs while a single DNA molecule translocates through the nanopore), blockage current ΔI = I_B − I_O, and translocation time of the molecule, t_duration. (d) Schematic showing (top) the chemical difference between cytosine and methylated cytosine; (middle) unmethylated versus a fully methylated CpG dinucleotide in dsDNA. Data traces of unmethylated- (bottom-left) and methylated-DLX1 (bottom-right) were recorded at 300 mV, showing similarity between both data traces. (e) Comparison of mDLX1 and uDLX1 transport. ΔI and τ_d plots as a function of applied voltage. τ_d and ΔI refers to the time constant and the blocking current respectively at each voltage. All points are the value of the fit with standard error. Second order of polynomial fit to ΔI and exponential fit to τ_d are also shown in short dash (black fits to uDLX1 and red to mDLX1). Each data are overlaid with over n = 1167 separate translocation events recorded per data point. Methylated and unmethylated fragments are indistinguishable. (f). τ_d (top) and ΔI (bottom) histograms for mDLX1 versus uDLX1 at 500 mV (n > 2153), showing similarity with τ_{d_mDLX1} = 0.124 ± 0.006 ms, τ_{d_uDLX1} = 0.135 ± 0.006 ms, ΔI__mDLX1 = −449.5 ± 5.4 pA and ΔI__uDLX1 = −440.8 ± 24.3 pA.

Figure 2. Differentiation of unmethylated DNA from mDLX1/MBD-1x complex.

(a) Structure of B-form dsDNA (left) and methylated DNA/MBD complex (right). A single MBD protein binds to the methylated CpG site on the major groove of dsDNA, occupying about 6 bps (PDB ID: 1IG4). (b) Top-down view: the cross-sectional diameter of the complex with a single bound MBD protein is ~5 nm. Multiple bound proteins along the DNA major groove increase complex diameter to ~7.6 nm. (c) Gel-shift assay showing the high affinity and specificity of MBD-1x for methylated but not unmethylated DNA. When increasing amounts of MBD-1x protein were incubated with uDLX1, no DNA-protein complex was formed (lanes 1–3), but when mDLX1 was included a robust, dose-dependent increase in mDLX1 - MBD-1x complex formation was observed (lanes 5–9) Lane 5 and 9 show 1:5 and 1:30 (mDLX:MBD-1x), respectively. Samples were fractionated on an 8% non-denaturing polyacrylamide gel and visualized using autoradiography. (d) Nanopore ionic current traces recorded in 600 mM KCl, pH 8.0 at 600 mV; uDLX1 events (left), mDLX1/MBD-1x events (right). (e) Characteristic translocation signatures for uDLX1 (bottom) versus the complex (top) through a ~12 nm pore. Scale bar is 10 nm in the TEM image. Qualitatively, the mDLX1/MBD-1x complex induces longer, deeper current blockades relative to uDLX1. (f) t_duration (left) and ΔI (right) histograms at 600 mV for uDLX1 (shown in blue − n = 857) and mDLX1 (shown in red − n = 197). Unmethylated DNA and the complex are clearly distinguishable. Exponential fits give time constants of τ_{d_uDLX1} = 0.103 ± 0.005 ms and τ_{d_mDLX1} = 1.43 ± 0.03 ms respectively.

Figure 3. Methylation quantification based on number of bound MBD-1x proteins.

MBD-1x protein was incubated with methylated DLX1 DNA at ratios of (a) 1:30, (b) 1:5 and (c) 1:1. Characteristic current signatures representing the mDLX1/MBD-1x complex (top) and unmethylated DLX1 DNA (bottom) through 9–10 nm diameter pores are shown. Current signature histogram of unmethylated DLX1 (black) and methylated DLX1-MBD-1x complex (red). Histogram was generated with peak current signature value of each event. (Inset) Scale bar is 10 nm in TEM images. (d) Translocation time histograms representing the mDLX1/MBD-1x complex and unmethylated DLX1 (inset). (e) Methylation Detection (left): Complexes formed with any ratio of MBD-1x can be discriminated from uDLX1 using blockage current alone (~3-fold increase in blockage current induced by the complex is seen). Methylation Quantification (right): Complexes formed with different ratios of protein can be differentiated based on the number of bound MBD-1x molecules. Time constants for the complexes are shown by the red circles: J_1:30 = 4.51 ± 0.48 ms, J_1:5 = 1.67 ± 0.17 ms and J_1:1 = 1.01 ± 0.09 ms. Corresponding time constants for uDLX1 were in the range of 0.107 – 0.184 ms MBD-1x on complexes were quantified with extended translocation duration. 1:1 complex showed ~7-fold prolonged translocation duration, 1:5 at ~12-fold and 1:30 at ~31-fold respectively than unmethylated DNA.

Figure 4. Molecular Dynamics (MD) simulations of methylated DNA/MBD complex through a nanopore.

Temporal MD snapshots showing translocation of 63 bp dsDNA with: (a) 3 bound MBD proteins through a 12 nm pore, (b) 3 bound MBD proteins through a 10 nm pore, (c) 1 bound MBD protein through a 9 nm pore. As pore size is reduced, hydrophobic interactions between the complex and the pore begin to dominate and can arrest the transport of the molecule through the pore. (d) Center of mass of the complex is shown distance vs. time. Smaller pore sizes can result in the trapping of the complex in the pore.