A team of researchers led by the National Institute of Standards and Technology has used a single protein nanopore to distinguish polymer molecules of different sizes. Their study — in effect, miniaturized mass spectrometry in solution — might help pave the way for nanopore-based DNA sequencing, according to one of the researchers involved.
The scientists, who published their work in the Proceedings of the National Academy of Sciences earlier this month, were able to distinguish polyethylene glycol molecules consisting of between 25 and 50 monomeric units. Measuring the degree and the duration of current blockades that these molecules caused while inside an alpha-hemolysin nanopore, the researchers could readily discern PEG molecules differing by as little as one monomeric unit, or 44 mass units. The technique also had sufficient resolving power to distinguish molecules that differ in mass by about 15 units.
Since submitting the paper, the group has been able to increase the resolving power by about 20 percent. “Hopefully soon, we will be able to determine the smallest difference between two molecules that can be obtained with this system,” said John Kasianowicz, leader of the nanobiotechnology project in the Semiconductor Electronics Division at NIST, who led the research.
Almost 11 years ago, Kasianowicz, together with researchers from Harvard University and the University of California, Santa Cruz, published a landmark paper in PNAS on characterizing single-stranded RNA and DNA molecules by threading them through an alpha-hemolysin channel, the first publication that introduced the concept of nanopore-based DNA sequencing.
In that paper, the researchers showed that they could drive DNA through the channel with an electric field, and were able to distinguish short polynucleotides from longer ones based on how long the molecules took to thread through the pore. In their most recent study, using neutral PEG molecules, the NIST team and their colleagues at the Universidade Federal de Pernambuco in Brazil showed that it is possible to accurately distinguish mass differences to much better than a single monomer.
In order to be able to sequence DNA based on such mass fingerprinting — similar to traditional capillary electrophoresis-based sequencing, but on a much smaller scale — the researchers have to improve the resolution even further, so they can distinguish between the four different DNA bases. However, this might be a difficult goal to achieve.
“The differences between the contributions that each of the four bases make to the total mass of a single-stranded DNA molecule inside a typical nanopore are not particularly great,” Kasianowicz said. “Therefore, to be able to determine the sequence of an individual polynucleotide based on the conductance blockade of a nanopore is a stretch. The idea is still interesting, but a practical solution is still elusive.”
“One challenge in looking at DNA ‘fingerprints’ is that DNA, which is charged, moves through the pore much more rapidly than PEG, giving much shorter blockades and less information in each signal,” Tom Butler, a graduate student in Jens Gundlach’s group at the University of Washington who is also working on protein nanopores, noted in an e-mail message. Also, since DNA monomers are much larger than PEG monomers, even relatively short DNA molecules already completely fill the protein pore, he added.
But Kasianowicz thinks there might be another way to sequence DNA using a nanopore: Each DNA molecule, he believes, is likely to produce a blockade pattern — or electronic “spectroscopic fingerprint” — that is not only characteristic of its mass, but also of the order of its bases.
The idea for these patterns came to Kasianowicz and his NIST colleague Vincent Stanford in the 1990s when they studied homopolymers of DNA traveling through a nanopore. Initially, they found that individual, homopolymer molecules of identical length caused different pore conductance blockades.
“I was initially concerned because the signals caused by seemingly simple [and identical] molecules appeared to be widely disparate,” Kasianowicz said. But he and Stanford found patterns hidden in the data that were characteristic for a specific type of homopolymer. Kasianowicz likened this to speech recognition software, which — if properly trained — can recognize a word even if it is spoken in different ways. “We looked at all the possible patterns caused by several kinds of DNA, and they were completely different,” he said. These patterns consist of the depths, number, and lifetimes of the blockade states, and the connectedness between these states, he explained.
“To be able to determine the sequence of an individual polynucleotide based on the conductance blockade of a nanopore is a stretch. The idea is still interesting, but a practical solution is still elusive.”
Their hope is that every DNA molecule type of interest will produce a characteristic conductance blockade pattern that a computer will be able to pick out. “Thus, even if you can’t floss an individual piece of DNA through the nanopore and obtain a base-by-base sequence, with a population of the same molecules, you might obtain blockade patterns that are specific for that particular DNA sequence, but not another,” Kasianowicz said, adding that researchers might be able to sequence strands consisting of “hundreds of bases” this way.
“Distinguishing between very similar populations of DNA molecules on the basis of their ‘electronic spectroscopic fingerprints’ is definitely possible,” said Butler. “Given the present demonstrated sensitivity of the system, this approach is much more realistic than measuring each individual base as it threads through the pore. The degree to which similar DNA molecules can be distinguished by their ‘fingerprints’ is a very important unanswered question in the field,” he added.
So far, Kasianowicz and his team have studied a number of 100-base-long DNA sequences, and have been able to distinguish between them based on the patterns they found. Some of this work was presented at an IEEE workshop on Genomic Signal Processing and Statistics in Baltimore, Md., three years ago and in a paper published in Analytical Chemistry in 2001.
But nanopore sequencing based on electronic "spectroscopic fingerprints" is probably still at least five to 10 years away, according to Kasianowicz. “Each year we and others are making reasonable leaps,” he said. “However, historically, it appears that about one to three years is needed to get something really worth talking about.”