This article has been updated from a previous version to clarify some of Mariam Ayub's comments.
The University of Oxford laboratory behind the alpha-hemolysin protein-based nanopore DNA sequencing strategy being developed by Oxford Nanopores Technologies has shown that it's possible to use similar nanopores for recognizing RNA bases.
As they reported online last week in the journal Nano Letters, Hagan Bayley and Mariam Ayub, a postdoctoral fellow in Bayley's University of Oxford chemistry lab, compared the signals associated with immobilized RNA oligonucleotides in wild type and modified alpha-hemolysin pores — a strategy similar to that used to discriminate between anchored DNA bases during preliminary studies that provided the foundation for alpha-hemolysin nanopore sequencing of DNA (IS 4/21/2009).
The proof-of-principle study demonstrated that modified alpha-hemolysin proteins with slightly larger than wild type pore diameters produced distinguishable signals for all four RNA bases, as well as three modified bases.
"Nanopore sequencing for RNA is still quite a new subject — most people have focused on DNA," Ayub told In Sequence. "The idea behind our paper was to really just put some light on whether we can actually sequence RNA in the same way as DNA."
The technique that they used for recognizing each RNA base in the pore was quite similar to that described for DNA sequencing with nanopores, Ayub explained, though RNA is somewhat less stable than DNA and prone to forming secondary structures.
"If you have a single strand of RNA, it has hairpins and loops in its structure, so it's not linear," she said. "When it passes through the pore, you have to unfold it."
That was not an issue for the stretches of single-stranded RNA that researchers were dealing with in the current study, which were relatively short and immobile. But it is a consideration when dealing with longer pieces of RNA.
Fortunately, Ayub explained, researchers believe that the same sorts of ratcheting methods that will be needed to translocate RNA strands through the nanopore at speeds suitable for sequencing should also iron out such kinks in the molecule.
To that end, she is currently looking at enzyme-based options for processively sequencing RNA molecules as they move through the alpha-hemolysin pore.
"Within this paper, what we show is … a static approach rather than an active approach," Ayub said. "This is really the first step."
"Something I'm working on right now is an active process," she added.
On that front, researchers suspect it may be possible to use an enzyme to help ratchet single-stranded RNA through a protein nanopore, similar to the approach that the University of Washington's Jens Gundlach and University of California, Santa Cruz, researcher Mark Akeson have described for DNA sequencing with the MspA pore protein (IS 3/27/2012).
Though she did not disclose the nature of the enzyme or enzymes being tested for RNA ratcheting, Ayub noted that either polymerase-type enzymes or exonucleases that lop off bases could serve as potential candidates for controlling the movement of DNA or RNA through nanopores.
Alpha-hemolysin, the pore-producing protein used to read immobilized RNA bases in the current study, has already been the subject of much development for DNA sequencing. In particular, Oxford Nanopore has been developing nanopore sequencing devices based on alpha-hemolysin research by Bayley, the company's founder and a member of its board of directors.
Earlier this year, Oxford Nanopore announced that it planned to introduce two commercial nanopore sequencing instruments this year, a scalable GridIon platform and a USB-stick sized MinIon sequencing system (IS 2/21/2012). Since then, the company has not disclosed further details of its commercialization timeline.
For the new study, Ayub and Bayley used RNA oligonucleotides tagged with biotin to characterize the current profiles that each RNA base would produce if it were to pass through alpha-hemolysin pores in a sequencing setting. By binding the biotin tag to streptavidin at the edge of the pore, they could anchor the end of the single-stranded RNA molecules that were coaxed into the pore using applied voltage.
"We looked at very short oligos — so short strands of 30 nucleotides — in these measurements," Ayub said.
These included so-called RNA homopolymers comprised of adenine, uracil, and cytosine repeats, as well as 30-nucleotide oligos in which a different base had been inserted into position nine with a run of poly-adenosine or poly-cytosine bases.
The latter approach was especially useful for looking at the pore's ability to pick up modified RNA bases and for determining the current associated with guanine, which is trickier to assemble into homopolymers than the other three RNA bases.
"You can't make a polymer of [guanine], either for DNA or RNA," Ayub said. "You can't go more than three bases" because it's an unstable structure and therefore difficult to synthesize.
To complement those analyses, the team also looked at the signals associated with each of the four unmodified and three modified RNA bases when they were placed at position nine of a similar DNA oligonucleotide comprised of cytosine — or "poly-C" — bases.
In the past, Ayub explained, Bayley's University of Oxford group and its collaborators had introduced mutations into the alpha-hemolysin protein that slightly modified the bottom end of the alpha-hemolysin pore barrel and changed the charge distribution at the constriction of the pore — changes aimed at improving the protein's performance for DNA sequencing applications, by making it easier to distinguish between DNA bases.
For the current study, those modified pores also proved useful for discriminating between current signals associated with RNA bases, Ayub noted. "As we modified these, we were able to get better signal-to-noise ratios [from longer residence times] as well as discrimination [due to the mutated pores]."
For the single-stranded RNA molecules tested, she explained, it was difficult to reliably distinguish one RNA base from the next using wild type alpha-hemolysin pores.
But, as previously shown for single-stranded DNA, when they tested modified pores with slightly larger pore diameters that allowed more current through, the investigators found that they got signals that allowed them to make out all four RNA bases.
They could also distinguish modified bases, including methylated forms of cytosine and adenosine and inosine, a modified RNA base formed through adenosine deamination.
"[W]e achieve sharply defined current distributions that enable clear discrimination of the four nucleobases, guanine, cytosine, adenine, and uracil, in RNA," Ayub and Bayley wrote. "Further, the modified bases, inosine. N6-methyladenosine, and n5-methylcytosine, can be distinguished."
A modified version of the alpha-hemolysin pore known as the NN pore offered improved resolution relative to the wild type pore. But another modified protein that produced the NNY pore was better still, Ayub and Bayley reported, producing even more distinct profiles for each RNA base.
So far the team has not had to make any appreciable changes to the system used to read and decipher the signals produced by RNA bases in the pore, Ayub said, noting that "there were no appreciable differences in terms of current noise and the results were comparable" to the group's previous work with immobilized DNA.
If it proves to be feasible to take the application a step further to sequence the RNA bases, she added, it should be possible to come up with an alpha-hemolysin-based nanopore sequencing system that can read either DNA or RNA molecules.
"It is possible to use the same pore — the same mutant of alpha-hemolysin — to observe similar results, either with DNA or RNA, provided that the system is optimized for that experiment," Ayub said.
But, she explained, the ability to swap back and forth between the two applications using the same nanopore platform may also depend somewhat on the strategy used to ratchet and ease the molecule through the pore as it is sequenced.
For instance, if a ratcheting enzyme and the pore protein are produced from a single construct, it won't be possible to easily swap between the enzymes for ratcheting DNA and RNA. On the other hand, if this system relies on enzymes that are temporarily docked on top of the pore, it would be easier to switch back and forth.
"If [the enzyme] is just docked on top, then obviously you could use the same setup and just add a different enzyme — whether you're looking at DNA or whether you're looking at RNA — once you have the conditions optimized," Ayub said.
Because the enzymes that recognize RNA and DNA often function optimally under different conditions, she noted that it will likely be necessary to switch out the buffers used between DNA and RNA sequencing setups as well.
For now, the focus is on finding that appropriate enzyme and demonstrating that modified alpha-hemolysin can be used in a system that recognizes not only static RNA bases but also those moving sequentially through the pore.
Based on work being done by other members of the research group, it appears to be possible to translocate "up to a couple of kilobases" of RNA through an alpha-hemolysin nanopore, Ayub noted, though more optimization is likely needed to sequence pieces of RNA of that length.
"The idea was, as a first step, to show that we can recognize individual bases of RNA just as well as we can for DNA," Ayub said. "The ultra-rapid sequencing of RNA will ultimately require an active process to control movement of the nucleic acid through the pore. Hence, it will be necessary to combine [alpha-hemolysin or an alternative protein nanopore with a processive RNA translocating enzyme, to ratchet RNA through the pore at a speed at which base identification is feasible."
In addition to coming up with strategies to ease RNA's trip through the pore, Ayub noted that there is ongoing research into ways of beefing up arrays so that signals from many, many alpha-hemolysin pores can be read simultaneously on a robust platform.
"When you're considering thousands and thousands of bases, in order to sequence them simultaneously, you need to have some kind of array where you have thousands of pores on the same platform and be able to sequence them at the same time," she said.
That might involve taking a hybrid approach (IS 12/21/2012) that incorporates solid-state nanopore features, Ayub explained, for instance, by finding ways to insert protein nanopores into channels in solid material.
"The idea is to basically combine the robustness of the solid-state nanopore with the precision of the alpha-hemolysin pore," she said. "Because you can fine tune the alpha-hemolysin pore to allow you to detect single bases, by combining it with a solid-state nanopore, you can begin to build a platform, which is robust."