Research teams at the University of California, Santa Cruz and the University of Washington have shown independently that biological nanopores can detect epigenetic modifications on individual bases within DNA strands.
In proof-of-principle experiments, published back-to-back online in PNAS last month, both groups showed that a modified MspA nanopore, coupled with phi29 DNA polymerase, is able to distinguish between unmodified cytosine, 5-methylcytosine, and 5-hydroxymethylcytosine in CpG dinucleotides within a DNA strand.
The results suggest that nanopores might be able to read a wide variety of epigenetic modifications directly off a DNA strand, which could someday have widespread applications in research as well as clinical studies.
Both teams — one led by Mark Akeson, a professor of biomolecular engineering at Santa Cruz, and the other by Jens Gundlach, a professor of physics at UW — used an experimental system described by Gundlach's group in 2012 (IS 3/27/2012), in which wildtype phi29 DNA polymerase feeds a single-stranded synthetic DNA oligomer through a modified MspA nanopore in single-nucleotide steps while it synthesizes DNA. It involves a blocking oligomer that prevents the polymerase from making DNA until the oligomer has entered the nanopore. Changes in the ion current through the pore are measured and correlated with the base sequence of the DNA.
The UW researchers compared ion current traces from single DNA molecules that had identical sequences but differed in their cytosine methylation status. They found that methylated C and hydroxymethylated C resulted in distinct current signatures over several bases, with methylation increasing and hydroxymethylation decreasing the current. Within a read, they found the detection efficiency to be about 98 percent for methylation and about 97 percent for hydroxymethylation.
The UC Santa Cruz group found that 5-methylcytosine and 5-hydroxymethylcytosine resulted in three consecutive ionic current states that differed between them and also depended on sequence context. They measured currents from about 3,300 DNA strands, representing 48 different oligomers, and found that the error rate for calling cytosine, methylcytosine, and hydroxmethylcytosine ranged from 1.7 percent to 12.2 percent for a single-pass read.
While these studies are not the first to show that nanopores can distinguish between methylated and unmethylated DNA in general, they are the first to pinpoint the exact location of the modification in a single strand of DNA, Gundlach told In Sequence.
Unlike bisulfite sequencing, a common method for reading cytosine methylation with single-base resolution, nanopore sequencing does not require the DNA to be chemically converted, which is cumbersome and damages the molecule, he said. "Our method is a lot more direct, ultimately faster, and … it's compatible with very low DNA quantities," he said.
Gundlach also believes their method is more sensitive than Pacific Biosciences', which also works with single DNA molecules and does not require the DNA to be modified. PacBio uses kinetic information from the polymerase to infer the methylation status of the template strand. "We believe our method is more direct and much more sensitive," Gundlach said.
In addition, it will probably be possible to use MspA to read other types of DNA modification — such as 8-oxo-guanine, thymine dimers, 5-carboxylcytosine, and 5-formylcytosine — and MspA has already been shown to record abasic sites with great sensitivity, he said.
"Virtually every modification that we've made on DNA over the years has been observable with one nanopore or another," said Akeson, adding that his group is working on expanding the list of epigenetic modifications that can be detected.
According to Akeson, other types of nanopores will likely also be capable of measuring DNA modifications, including biological and solid-state nanopores. In addition, researchers will use other types of "motor" enzymes, besides phi29 polymerase, to control the movement of DNA through the pore, and solid-state techniques might eventually be able to control DNA movement without an enzyme. "This is just the beginning," he said.
"Everyone is working very hard on new motors, improving pores, even though MspA is phenomenally sensitive," Gundlach said. "It's a technique that right now works with MspA. Of course for industrial implementation, you need to improve here and there, [and] especially work on the delivery of DNA to the pore."
Both groups now plan to apply nanopore methylation mapping to biological DNA rather than synthetic oligos.
The main challenge for doing so is the fact that genomic DNA is less concentrated than synthetic DNA, making it less likely to hit a nanopore. To get around that, Akeson's group is drawing on a technique developed and patented by Oxford Nanopore Technologies to concentrate the DNA on the surface of the membrane that holds the nanopores.
Tethering the DNA to the membrane can be achieved by modifying either the membrane or the DNA. "The main benefit of tethering is a drastic reduction in the concentration of analyte needed to obtain a useful experimental output," a spokesperson from Oxford Nanopore told In Sequence. "Essentially the nanopore is 'searching' in 2D rather than 3D for an analyte." As a result, the signal increases from 1,000-fold to 10,000-fold for a given amount of DNA.
Akeson said he plans to study the epigenetics of well-characterized biological DNA, including Arabidopsis and mouse DNA.
In addition, his group is working on ways to re-read the same DNA strand several times over in the same nanopore, using a currently confidential strategy that "we already know can work." Re-reading the same molecule would increase the accuracy for a particular methylation site, he said.
Companies working on commercial nanopore sequencing systems are likely to incorporate epigenetic modification detection in their offerings.
Akeson is a consultant to Oxford Nanopore Technologies, and the company continues to work with him and his team "on this exciting work and more," according to the firm's spokesperson.
Illumina, meanwhile, recently took an exclusive license to the MspA sequencing technology developed by Gundlach and his colleague Michael Niederweis at the University of Alabama (GWDN 10/15/2013), but Gundlach declined to comment on how the company plans to commercialize it.
Nanopore sequencing is "approaching a state where various groups have shown that all the components can work," Akeson said. "Now, it's going to be much more crowded; a lot more people are going to be doing nanopore sequencing."