By Monica Heger
Researchers have demonstrated that an alpha-hemolysin nanopore can identify epigenetic modifications of DNA, suggesting that eventually, nanopores will be able to perform single-molecule methylation sequencing of a DNA strand — an advance over current methylation sequencing methods, such as bisulfite sequencing, which require the DNA to be modified or amplified.
The study was published last week in Chemical Communications by Hagan Bayley's group at the University of Oxford. Bayley is a co-founder of Oxford Nanopore, which supports much of his research (IS 9/14/2010).
The work builds on a previous study, published this August in Nano Letters, where Bayley's group showed that it could enhance one of the three base recognition sites in the alpha-hemolysin pore to maximize the signal from that site. The group was able to demonstrate distinct currents for each of the four different bases after modifying the pore.
"We mutated the protein at a particular position," which increased the signal generated when a base was in the recognition site, Bayley told In Sequence last week.
By adding different amino acid side chains near the recognition site, the team was able to either improve or diminish the pore's ability to distinguish between bases. The researchers tested different configurations on a single-stranded chain of DNA containing one adenine base among numerous cytosine bases, and found that, in general, the addition of hydrophobic amino acids increased the ability to distinguish between bases.
"The differences in the ability to recognize nucleobases exhibited by [the modified and unmodified] pores are remarkable, given that there is only a single amino acid change, and provide evidence that it is possible to either greatly weaken or significantly improve nucleobase recognitions by mutagenesis," the authors wrote in the Nano Letters paper.
The Chemical Communications study adds to that work by testing the same mutated pore on methylated and hydroxymethylated DNA. A single-stranded piece of DNA containing one methylated CpG island in a string of cytosine bases is suspended in the pore. When voltage is applied, the methylated cytosine produces a current distinct from that of any of the other nucleobases. Similarly, a hydroxymethylated cytosine produces yet another current when suspended at the recognition site in the pore.
"We showed we were able to obtain different current levels" for all four bases and the methylated and hydroxymethylated cytosines, Bayley said. "It should be possible to pick those up in a nanopore sequencer, which would enable single-molecule sequencing of unamplified DNA."
The work builds on an earlier study, in which the Oxford team showed that the alpha-hemolysin can distinguish between different nucleotides, including methylated cyotosine (IS 2/24/2009), that are not in the context of a DNA strand.
There is still a ways to go before the nanopore can be realized in a device, however. For instance, while Bayley's team was able to identify six distinct signals, the signals were elicited from single bases placed within a specific and known sequence, and while the DNA was suspended in the pore. One unknown, said Bayley, is the effect on the signal of different bases adjacent to the base being read.
[ pagebreak ]
"If your reading head is not particularly sharp, then the [reading for the] base at position 150 might be affected by what's at position 149 and 151," Bayley said. "The actual current level might be modified by neighboring bases."
This is one reason it could be advantageous to use a pore like alpha-hemolysin. The alpha-hemolysin pore has three different recognition sites that could be employed to read bases. While Bayley's team has demonstrated that one of those sites can be enhanced so that current to read the base is only coming from that particular site, he says it would also be possible to employ a second reading site. The additional information provided by a second reading site could make it easier to distinguish the individual bases, because researchers would get two different signals for each base, resulting in higher accuracy, he said.
However, interpreting current from all three recognition sites would likely prove too complicated, he said. Aside from mutating the protein to enhance a specific recognition site's signal, Bayley said his team is also working on modifying the pores to reduce the signal.
A group from the University of Washington recently demonstrated that a different pore, made of Mycobacterium smegmatis porin A, could also distinguish individual bases in a single strand of DNA. The MspA pore is simpler than the alpha-hemolysin pore and only contains one recognition site, which the researchers said would give it an advantage (IS 8/24/2010).
Jens Gundlach, a University of Washington biophysicist who led the MspA research, said that the site where the bases are read is nearly identical in the two pores. The primary difference is that the MspA pore only has one reading site, while the alpha-hemolysin pore has three different sites. As a result, the current differences between the four bases are more pronounced in the MspA pore, he said.
While Bayley said that having multiple reading sites could be an advantage because there would be two chances to read each base, Gundlach argued that two reading sites would make the informatics more complicated because it "reduces the current dispersion between the bases."
On the other hand, Gundlach noted that the ability of the alpha-hemolysin pore to distinguish methylation would be a huge advantage. "That is a really big deal," he said.
Gundlach noted that current methylation sequencing methods require researchers to first convert cytosine to uracil, leaving lots of opportunity for error. He added that Bayley's group observed an unusually large current difference between the methylated and unmethylated cytosines, compared to what was observed between the individual bases. "It's surprising, but really nice," he said.
Mark Akeson, a professor of biomolecular engineering at the University of California, Santa Cruz, who also works on protein nanopores, added that Bayley's group demonstrated the ability to precisely engineer a protein nanopore. Akeson is also a member of the technical advisory board of Oxford Nanopore Technologies.
"The protein engineering is remarkable. A single amino acid substitution in the protein can give you structural and functional differences at the Angstrom scale," he said, which is difficult to do in technology such as solid-state nanopores.
He added that at this point, it's unclear whether the multiple recognition sites of the alpha-hemolysin pore would prove to be advantageous. "It remains to be seen how finely the pore can be engineered."
Akeson's group recently demonstrated a technique that uses a polymerase to slow down DNA passing through a nanopore — another major hurdle of nanopore sequencing (IS 9/28/2010).
Bayley said that his team is now working on methods to slow down DNA, too, and has also used polymerases to help with the process. "For the time being, we're more out of the game of base identification and are focused on slowing down DNA translocation," he said.
Another goal is making the approach higher-throughput by enabling many nanopores to read DNA simultaneously, he said.