Researchers at the University of Oxford have come one step closer to nanopore sequencing after showing that alpha-hemolysin can recognize all four bases of DNA in a strand immobilized inside the pore, regardless of the surrounding bases.
In their study, appearing online in PNAS this week, the researchers showed that the tunnel-shaped half of alpha-hemolysin contains three base-recognition sites that can identify single bases in a DNA strand anchored inside. However, reading off the bases from moving DNA remains a challenge.
The work is noteworthy because the researchers showed they can "identify all four bases of DNA in a heteropolymeric background" of DNA, senior author Hagan Bayley, a professor of chemistry at Oxford, told In Sequence. "That's not been done before."
He and his team threaded single-stranded DNA oligonucleotides through alpha-hemolysin pores and stuck them inside by putting a streptavidin plug at one end of the strand and applying a voltage that pulls at the other end. They studied two different types of alpha-hemolysin: the wild type protein and a mutant version that contained two altered amino acids at the constriction site in the center.
By measuring how oligonucleotides with different base compositions blocked the ion current going through the pore, they figured out that the bottom half of the protein — a tunnel known as the beta-barrel — contains three distinct base-recognition sites.
A 2005 study by researchers at the Scripps Research Institute, in collaboration with Bayley, revealed only one such recognition site near the end of the tunnel that was able to distinguish between single bases. However, that study used a DNA hairpin to hold the molecule in the pore, which might have interfered with the recognition.
In the current study, further experiments with poly-C oligos that contain a single A base, each at a different position within the oligo, showed that all three recognition sites can tell A bases from C bases.
In addition, both the wild type and the mutant pore were able to distinguish single nucleotides of all four bases when the nucleotides were positioned near the second recognition site. The mutant pore was also able to do that when the bases were placed near the third recognition site.
Finally, the mutant protein could discriminate between all four bases, located near the second recognition site, in a heteropolymeric oligonucleotide, suggesting that neighboring bases do not interfere with the recognition.
Despite their success in reading single bases within the context of an immobilized oligo, challenges remain for doing the same with DNA traveling through the pore.
For a start, the researchers caution in their paper that even the most promising of the three recognition sites "might be too 'blunt' to distinguish all four bases in a diversity of contexts." One solution might be to "sharpen" one site and "blunt" the others through site-specific modifications to the protein, they suggest.
"Ideally, you want a single recognition site that just gives you the right answer — you just run it past a recognition site and read off the sequence," Bayley said, adding that his group is currently trying to engineer the sites. "I think there are lots of things to explore there, and I don't think there is any reason to believe that we can't make things better."
However, he cautioned, "it may not be possible to build a single, perfect recognition site." Instead, he said, it may be possible to combine signals from two or all three sites, although "it would be a little more complicated, conceptually, to extract that information."
[ pagebreak ]
Jens Gundlach, a professor at the University of Washington who is working on a different protein pore for DNA sequencing, called the Oxford researchers' results "very interesting" but pointed out that the differences in current between different nucleotides is relatively small.
Ultimately, researchers will want to read DNA moving through the pore, rather than immobilized DNA, "and with small current differences as they have shown here, that is going to be a little bit tricky," he said.
Gundlach's group is developing MspA for sequencing, another pore-forming protein, and he said the simpler geometry of that pore might give it an edge over alpha-hemolysin for recognizing single nucleotides within a DNA strand (see In Sequence 1/13/2009).
Other scientists wonder what the Oxford researchers' results might mean for moving DNA. The work "is important in that it demonstrates the localization of the recognition sites in single-stranded DNA, but we do not know how this will apply when attempting to sequence DNA strands that are actively translocating the pore," Andy Hibbs, founder and leader of Electronic Bio Sciences, a San Diego-based company, said via e-mail. The presence of multiple recognition sites "may turn out to be a hindrance that must be removed by appropriate pore modification," he added.
EBS has made "high-bandwidth, low-noise measurements" of single-stranded DNA moving through alpha-hemolysin, he said, and found that "there is likely a single region at the constriction of the pore that dominates the overall blocking level during translocation." The Oxford researchers did not resolve differences in that region, he said, and "there is no clear disagreement between our results and those now reported."
Good base recognition is only one challenge researchers need to overcome to make nanopore strand sequencing feasible. Another requirement is to slow down the DNA, which could be achieved in several ways.
"Slowing down the DNA is something we all have to do," Gundlach said. One promising approach, he said, might be to use magnetic tweezers (see other feature in this issue). According to Bayley, another way would require mutating the pore so that DNA binds more tightly to it. However, "we have not been so far successful in doing that," he said.
Another method Bayley's lab is exploring is to add a nucleic acid-binding protein that helps move the DNA through the pore more slowly. Such an enzyme could also help resolve another problem, counting several bases of the same type that move through the pore in sequence. "If you have an enzyme that's ratcheting the DNA one [base] at time, then you have a built-in counting mechanism," Bayley said.
Alternatively, he said, scientists hope "that there is a change in the current between each base … but we have no real evidence that we can get that type of spatial resolution."
Given these challenges, nanopore strand sequencing might not be the first nanopore-based sequencing technology to be commercialized. Oxford Nanopore Technologies, which has licensed Bayley's intellectual property, currently focuses on exonuclease nanopore sequencing, where alpha-hemolysin recognizes individual bases that are cleaved off by an exonuclease (see In Sequence 2/24/2009).
However, the company's IP also covers strand sequencing, and technologies it is developing now for exonuclease nanopore sequencing — such as nanopore array chips, an instrument, and software — might also benefit later generations of nanopore sequencing technology, according to a company spokesperson.