NEW YORK (GenomeWeb News) – Researchers from Pacific Biosciences reported today that they have developed a single-molecule, real-time sequencing-based method for directly detecting DNA methylation patterns without bisulfite or other treatments that alter specific nucleotides.
The proof-of-principle study appears in the early, online version of Nature Methods and describes how changes in the kinetics of the polymerase enzyme employed in the SMRT sequencing process can be used to spot methylation in synthetic DNA templates and DNA from a natural system.
Because methylation affects both the time it takes to load each nucleotide on the DNA polymerase used in SMRT and the time it takes for the enzyme to process the nucleotides, senior author Stephen Turner, Pacific Biosciences' founder and chief technology officer, told GenomeWeb Daily News, the team was able to assess the methylation status of each nucleotide by looking at how long each fluorescently labeled base interacts with the polymerase as well as the time between these pulses.
Those involved say the findings may open the door to simultaneous sequencing and methylation detection — and hint at the possibility of using SMRT sequencing to find other kinds of DNA modifications in the future.
Pacific Biosciences' SMRT sequencing approach relies on measuring fluorescence as a DNA polymerase enzyme is assembling labeled nucleotides into complementary DNA strands within tiny structures called zero-mode waveguides.
While characterizing this incorporation process, Turner explained, the researchers noticed variation in polymerase kinetics depending on which of the four bases was being processed. Based on these findings, they suspected DNA modifications such as methylation, which slightly alter the chemical structure of DNA, would also produce a kinetic change.
"[W]e hypothesized that methylated bases in a DNA template might be detected directly on the principle that their presence affects polymerase kinetics during SMRT sequencing," Turner and his co-authors wrote. "Curiously, to our knowledge the kinetics of nucleotide incorporation opposite methylated templates have not been studied previously … despite evidence that other types of modified nucleotides alter DNA polymerase kinetics."
To test this, the researchers started by sequencing a handful of synthetic DNA templates that were either unmethylated or contained bases with N6-methyladenine, 5-methylcytosine, or 5-hydroxymethylcytosine methylation.
In a follow-up to these initial experiments, they also used principal component analysis to compare the signals generated by bases with cytosine, 5-methylcytosine, and 5-hydroxymethylcytosine methylation. Indeed, the researchers found that they could not only detect methylation, but also distinguish between the various types of methylation tested.
Because methylation affects DNA's secondary structure, changing its pitch, twist, and rise, Turner said, this modification can lead to as much as a seven-fold variation in enzyme kinetics. These changes not only occur at the methylated base itself, he added, but lead to changes over a set of bases that the polymerase enzyme encounters as it deals with the methylated base.
"There is not just a signal, but a whole signature," Turner explained.
And this direct methylation detection was not limited to synthetic DNA templates. When the researchers sequenced Caenorhabditis elegans fosmid DNA from a strain of Escherichia coli containing the DNA adenine methyltransferase enzyme, they were again able to distinguish between methylated bases — in this case methylated adenine bases — and unmethylated controls.
Based on their findings so far, Turner said the researchers believe it will be possible to simultaneously re-sequence genomic DNA and generate data for assessing methylation status of the sequence. He noted that the team has tried using the method to assess methylation patterns in the human genome, but said it's too early to discuss the results of these experiments.
And while the current paper doesn't test whether it's possible to get direct methylation data in the context of de novo sequencing, Turner said this also appears to be a promising future direction for the approach.
In addition, the team is exploring kinetic patterns associated with other sorts of DNA modifications, including DNA damage, cancer-related changes, and more. And so far, Turner said, every type of modification tested seems to produce some detectable change in enzyme kinetics during SMRT sequencing.
Turner predicts that the methylation detection project will come out of the exploratory research phase and enter development relatively soon. As things stand at the moment, he added, Pac Bio hopes to offer a software upgrade for detecting methylation to users by the middle of next year.
In the meantime, those involved in the project are also trying to determine whether slight modifications to the polymerase enzyme might enhance their ability to detect methylation and other modifications.