By Monica Heger
Efforts to sequence DNA by threading it through protein-based nanopores have traditionally relied on one protein: alpha-hemolysin. But researchers from the University of Washington have diverged from that route, demonstrating in a recent proof-of-principle study that engineered Mycobacterium smegmatis porin A could yield a superior nanopore.
In a study scheduled for publication in the Proceedings of the National Academy of Sciences, the U Wash researchers showed that the engineered protein pore could distinguish between single nucleotidesas DNA passed through the protein nanopore that had double-stranded DNA hairpins inserted between each base to slow down the rate of translocation.
"It's a proof of principle that nanopore sequencing is going to work. Now it's just a matter of fine-tuning the method," said David Deamer, a chemist at the University of California, Santa Cruz, who was not affiliated with the study but who also works with protein nanopores.
"This is very impressive work that has, for the first time, generated real experimental data that mirrors the idealized cartoon where the nanopore current flips between four steady current levels, one corresponding to each base," said Ken Healy, a physicist at the University College Cork in Ireland, and part of a University of Pennsylvania team that recently demonstrated DNA translocation through a graphene nanopore (IS 8/3/2010).
The University of Washington team presented preliminary results of this work at the National Human Genome Research Institute's Advanced DNA Sequencing Technology Development meeting in Chapel Hill, NC, in March (IS 3/16/2010), but this is the first published demonstration of the research, which was funded by a grant as part of NHGRI's $1,000 genome award program.
The key to the team's ability to demonstrate single-base resolution is in the shape of the protein pore, according to the researchers. Nanopores made from alpha-hemolysin are typically around 5 nanometers long, so many bases are present in the pore at any given time, making it difficult to distinguish the signal of one single base, said Jens Gundlach, a biophysicist at the University of Washington who led the research. The MspA pore is much smaller — about 0.5 nanometers long, and about 1.2 nanometers in diameter — forming a constriction zone that has the "perfect shape for sequencing," Gundlach said.
That perfect shape, said Deamer, is a funnel. By contrast, alpha-hemolysin pores have more of a mushroom shape. So where the funnel-like MspA pore narrows down considerably at the site where the individual bases are read, the mushroom-shaped alpha- hemolysin pores allow around 10 bases to reside in the stem of the mushroom, all of which can contribute to the ionic current. However, researchers from the University of Oxford demonstrated last year that the stem contains three distinct base recognition sites that can identify all four bases of DNA (IS 4/21/2009).
Another advantage of the MspA pore is that it has a lower signal-to-noise ratio than alpha-hemolysin, according to Gundlach. He said the team observed around a 50 picoamp current difference between each base, approximately 10 times greater than the difference between each base observed with alpha-hemolysin pores.
The second trick was the team's tactic for slowing down the rate of translocation by inserting a double-stranded DNA hairpin between each single base of unknown sequence. The DNA hairpin is too thick to fit through the pore, so translocation is paused, giving the researchers enough time to read the signal from the single base trapped in the pore. After about 10 milliseconds, the voltage causes the hairpin to dissociate, and translocation again proceeds until the next hairpin is reached.
After determining the current signatures associated with each individual base, the researchers constructed a known five-base sequence of single-stranded DNA, with the same 14-base DNA duplex in between each base. When they ran the construct through the pore, they were able to distinguish each individual base. They then tested it on an unknown sequence of five bases, and were again able to accurately read each of the individual bases.
"The double-stranded DNA inserts a pause in between every nucleotide, and during this pause we have enough time to read each nucleotide," Gundlach said. "The read lengths we have are rather modest, but it represents a core of a potential sequencing technology, and it shows that nanopores in general have the potential to sequence DNA."
One drawback of the current method is that the DNA has to first be modified before it can be sequenced, Gundlach said. He said his team is now working on ways to slow down translocation without first modifying the DNA with DNA hairpins.
He said they will also work to improve the pore itself. The modified MspA pore was the first to allow DNA translocation, but Gundlach said that creating different mutations could improve its function even more.
Gundlach said that he and his colleagues have applied for patents on the method, and that they have been talking to potential industry partners, but he could not disclose any details.
Proteins vs. Solid-State
Gundlach said that protein nanopores are much more reproducible and simpler than solid-state nanopores.
"Every pore that we produce looks exactly the same," he said. In addition, researchers can take advantage of the ability of the protein nanopore to interact with the DNA through hydrogen bonds, amino acid bonding, and hydrophobic and hydrophilic interactions. Those interactions "all help generate the distinct current patterns."
However, one disadvantage of the protein nanopores is that they are not as stable as solid-state nanopores. "They are in a lipid membrane and they can be used for a few hours, sometimes for days, but not indefinitely," Deamer said. "Right now, we think of them as disposables."
Also, the reproducibility of protein nanopores can work to their disadvantage, said Slaven Garaj, whose team from Harvard University recently demonstrated DNA translocation through a one-atom-thick graphene nanopore in a paper published in Nature. "Protein nanopores are not versatile. They are always the same, which is both a strength and a weakness," said Garaj.
Garaj explained that graphene and other solid-state nanopores are versatile, allowing researchers to explore different methods for reading the bases, such as electrical readouts or optical detection. For instance, he said his team is now exploring graphene's unique electrical properties so that it can construct a graphene pore that would produce a different electrical signal as each individual base moved through the pore.
Several researchers believe that both solid-state and protein nanopores will have a place in future applications of DNA sequencing.
"The protein pores will be the next stage of sequencing and the solid-state pores will follow as we get better control over their size and the way they interact with DNA," Deamer said.
Deamer said that protein nanopore sequencing devices would reach the market first, and estimated that a device would be in the field within the next few years, noting that the switch from second-generation sequencers to nanopore sequencers will be "like the conversion from mainframe computers to personal computers."