Speed-reading DNA

Funneling DNA through nanopores enables fast, direct sequencing

Over the last decade, scientists' ability to rapidly read the genetic codes of organisms from humans to flatworms has increased dramatically. However, as with other technologies today, faster is often not fast enough. The desire to further accelerate DNA processing led Aleksei Aksimentiev to investigate using nanopores—the tiny openings cells use to pass material through their outer membranes—as electrical portals to read the genetic code of DNA directly as it passes through.

In this molecular dynamics simulation, a single strand of DNA is transported through a nanopore.
Simulation: Anthony Ho and Aleksei Aksimentiev, University of Illinois
In this molecular dynamics simulation, a single strand of DNA is transported through a nanopore. Simulation: Anthony Ho and Aleksei Aksimentiev, University of Illinois.

Aksimentiev, an associate professor of physics at the University of Illinois-Urbana-Champaign, explains the heart of a system he and his team have devised—a single nanopore mounted in a very thin membrane that divides two compartments filled with saltwater. When DNA is added to one of the compartments, an electric field drives the long, string-like molecules from one compartment to the other through the nanopore.

DNA is made up of four different chemical bases. As they pass through the nanopore, the size, shape and orientation of each base affects the current flowing through the pore differently. This creates a distinctive electric signature that allows each base to be identified and recorded.

Cheaper, faster

Sequencing DNA using this technique has several advantages over traditional methods. "The critical one," Aksimentiev says, "is the length of the string one can read." Usually when people try to read the sequence of DNA bases, they cut the molecular strings into small pieces because other reading techniques are limited to relatively short molecules. Afterward, researchers make assumptions about how the resulting sequences of bases should be reassembled. This process is costly, timeconsuming and sometimes inaccurate if the molecular strings are very short. "Using nanopore technology, we can read DNA sequences up to hundreds of thousands of bases long," he says. "This isn't possible by any other means, and it yields more accurate results."

The second advantage of the system is that it reads the sequence of bases that make up the DNA directly from the strand. Other techniques often involve attaching chemical labels to each base and then looking for the labels. "Our method involves passing DNA through a nanopore and determining the identity of each nucleotide based on its shape and electric signature," Aksemientev says, "so there is no need to modify or mark the bases prior to detection. We now are focusing on improving our ability to identify individual bases."

Better sequencing through simulation

Taking full advantages of ORNL's resources, Aksimentiev uses molecular dynamics simulations on the laboratory's Jaguar supercomputer to analyze his results and refine his team's methods. These simulations mimic the design of the system Aksimentiev and his colleagues have been using and help them answer a few key questions: "First," he says, "we don't know exactly why the electrical current moving through the pore is affected by the DNA bases. We know larger ones block the current less than the smaller ones, which is counterintuitive, so there must be some other physical phenomenon involved." Aksimentiev is using Jaguar to determine the position of every atom in the system. He expects simulations will reveal explanations for changes in current flow. "Experimentally, it is not possible to tell how the DNA is positioned as it passes through the pore," he explains. "Our preliminary results with the simulation suggest that it's not just the sequence of the DNA, but also its conformation, or structure—and changes in its conformation—within the pore, that contribute to changes in the current."

The challenge of speed

"Second," Aksimentiev says, "in our system we're using now, the DNA moves too quickly through the pore, making it hard to read its sequence. We are trying to genetically modify the channel so it interacts more strongly with the DNA, slowing it down as it passes through the pore, so its sequence can be detected more accurately." Simulation was used to design a previous round of genetic modifications aimed at putting the brakes on DNA moving through the pore. Aksimentiev notes that those changes were implemented by his colleagues in the lab and had the desired effect, but more slowing is needed. Even with this reduction in speed, Aksimentiev says the system is still fast enough, in theory, to sequence the entire human genome in an hour.

Finally, Aksimentiev explains that the current passing through the pore is affected by not just a single base, but by a series of three—the base passing through the pore, as well as the one ahead of it and the one behind it. He is using simulation to help determine whether the system can identify these "DNA triplets" by looking at their electric signature or, better yet, whether his colleagues can genetically modify the channel to be sensitive to a single base at a time. "Once we understand which of the three nucleotides passing through the pore provides the dominant contribution to the current," he says, "then we can suggest modifications to the pore that my colleagues can implement. That should improve the accuracy of the detection. We're making progress," he says, "but we will need to do more simulation to pinpoint the origin of each component of the electronic signature."

Next steps

The near-term goals for Aksimentiev and his colleagues include working with an industrial partner to commercialize nanopore sequencing technology. Their goal is to enable any researcher working in a laboratory setting to analyze genetic sequences quickly and easily. He and his colleagues are also trying to develop membranes and pores that last longer than the biologically based materials they're currently using. He says a synthetic membrane and pore combination would integrate more easily with electronics, but there are manufacturing challenges to overcome.

In the longer term, Aksimentiev anticipates broadening application of the technology to include identifying proteins. "Proteins are also linear polymers," he says, "but instead of just four bases, they have 20 different amino acids to describe their chemical makeup; and they don't carry a uniform electrical charge, so it is more of a technological challenge to move them through a nanopore." More importantly, however, there aren't nearly as many sequencing techniques for proteins as there are for DNA, so the impact of successfully applying this approach to proteins could be even greater."

Aksimentiev notes that computer simulations run on Jaguar are by far the best tools for exploring complex interactions like those involved in protein sequencing. To date, only a few protein sequencing experiments that involved nanopores have provided useful data, but the cause of the problem isn't clear. "We need help in understanding how to move proteins through the pore in a controlled way," he says. "That's something that Jaguar can clearly help to explain."— Jim Pearce