BALTIMORE – Researchers from the University of Bergen in Norway and their collaborators have developed a new approach to profile RNA structure at the single-molecule level using nanopore sequencing.
The method, named single-molecule structure sequencing (SMS-seq) and described in a study published in Nucleic Acids Research in September, combines a new chemical structural probing scheme with direct RNA nanopore sequencing, expanding the toolkit for studying RNA structure at the single-molecule level without the need for reverse transcription and PCR.
"RNA structure is extremely important in many molecules," said Eivind Valen, a computational biology professor at the University of Bergen and the corresponding author of the paper. "What we wanted to figure out was how much information we can get out of the single [RNA] molecules."
Previously, RNA structures were mostly detected by structure-dependent chemical probing followed by short-read sequencing. In these experiments, RNA molecules are tagged with chemicals that can modify the ribose 2′-hydroxyl group — such as selective 2′-hydroxyl acylation analyzed by primer extension (SHAPE) reagents — or chemicals that can tag the nucleobase, including dimethyl sulfate (DMS).
During cDNA synthesis, the chemically modified RNA can lead to reverse transcription termination or base misincorporation, which can then be captured through sequencing. Such information can help scientists infer RNA structures computationally.
However, as the study pointed out, there are several limitations associated with the short-read sequencing methods. For instance, these approaches cannot effectively achieve full-length structural profiling of larger RNA molecules such as messenger RNAs and ribosomal RNAs due to read-length constraint. Additionally, the need for PCR amplification means these methods generally cannot capture native modifications on the RNA molecules and cannot analyze RNA structure at the single-molecule level.
To overcome these bottlenecks, Valen and his team sought to combine RNA structure probing with direct RNA nanopore sequencing. The researchers initially tried using the SHAPE reagents and DMS, however, they found both chemical probing schemes resulted in "low-quality, truncated, and unmappable sequences," according to the paper.
Because the SHAPE reagents modify the RNA backbone and are "quite bulky," the modified RNA molecules didn't pass through the pore very efficiently, Valen said. "Ideally, you would like something that modifies all open bases, doesn't degrade the RNA, and can pass through and be detected with the nanopore," he added.
To that end, the researchers adopted a new probing scheme using diethyl pyrocarbonate (DEPC) — a chemical that carbethoxylates unpaired adenosines within the RNA and, according to Valen, introduces "a very high modification level" without causing significant RNA breakdown.
To achieve SMS-seq, RNA molecules were treated with DEPC, followed by direct RNA sequencing using the MinIon sequencer from Oxford Nanopore Technologies. As the RNA translocated through the nanopore, the researchers looked for deviations in the current signals between the control and modified samples to detect structure-dependent modifications on an RNA molecule. From that, they derived the DEPC modification sites for adenosine using a computational software called Tombo.
Overall, the study showed that SMS-seq can help detect transient and stable RNA structures, secondary and tertiary structural elements, structural dependencies of bases, and mRNA structural features.
In particular, when applying the method to a synthesized RNA hairpin with known structure, the researchers found that SMS-seq can help reveal the co-dependence of bases and the conformational heterogeneity of RNA structures.
The authors also tested the method on riboswitches, a regulatory segment of messenger RNA molecules that controls transcription activity by binding to small molecules. The results showed that SMS-seq can help reveal features of structural dynamics, such as conformational changes, short- and long-range structural interactions, and tertiary RNA interactions, when performed on two riboswitches, thiamine pyrophosphate (TPP) and Fusobacterium nucleatum Flavin mononucleotide (FMN).
Despite the method’s promises, SMS-seq has some current limitations. For one, DEPC can only be used in in vitro and therefore cannot be applied to live cells, Valen said. In addition, it can only modify adenosines. The modification rate for DEPC in the study was about 25 percent in the synthetic RNA samples, but "ideally, you would like to modify everything that's open, so you can get a perfect landscape," he said. "We're quite far from that."
"This is a further addition to the methods that we can use to chemically probe RNA structure using direct RNA sequencing with nanopore," said Winston Timp, a biomedical engineering professor at Johns Hopkins University. "Usually when we say RNA sequencing, we're doing cDNA sequencing … with direct RNA sequencing, you're actually looking at the RNA molecule itself."
Prior to SMS-seq, researchers including Timp had started to explore using nanopore sequencing to achieve direct detection of RNA modifications and structures. In a study published earlier this year, Timp and his collaborators described nanoSHAPE, an RNA structure probing method that pairs a new SHAPE reagent called acetylimidazole (AcIm) and direct RNA nanopore sequencing to achieve full-length structure profiling of long RNAs.
In addition, in 2020, researchers from the Genome Institute of Singapore published a method named RNA structure analysis using nanopore sequencing (PORE-cupine). In that approach, SHAPE reagent 2-methylnicotinic acid imidazolide azide (NAI-N3) was used for structure probing, in conjunction with direct long-read RNA sequencing and machine learning to help detect RNA structures.
Compared with the SHAPE reagents, one advantage of DEPC is that it can be easier to work with, Timp noted. "It's a very common chemical that is cheap and easy to buy and [can produce] lots of modifications," he said.
However, echoing Valen’s point, Timp said the downside of DEPC is that the reagent can only modify adenosines and only works in vitro. "If you have the RNA out of the cell, DEPC is probably better," he said. "If you have the RNA still in the cell, you can't really use DEPC."
Despite the potential advantages of direct RNA nanopore sequencing, Timp said, the yield for the technology is currently still "substantially lower" than for cDNA sequencing, which can be a bottleneck for the technology.
Additionally, he said Oxford Nanopore’s direct RNA sequencing feature is, as of now, only compatible with its 9.4 flow cells and is not yet available on the newer 10.4 flow cells. As the company plans to phase out the older flow cells down the road, "one concern I have, and I encourage Oxford Nanopore to do this, is that they continue to develop library prep for direct RNA sequencing on the newer 10.4 [flow cells]," Timp said.
Moving forward, Valen said he hopes the SMS-seq method can continue to be improved. "We did try to take this to the next step and try to fold individual molecules," he noted. "I'm not sure we are really ready for that step yet." That said, he hopes for further improvements in RNA probing reagents that can sufficiently tag RNA molecules for structure determination without compromising their integrity or hindering nanopore sequencing.
"I think, as these methods improve, particularly as the base calling of modified bases in nanopore improves and the chemicals that they use, this is going to be a very valuable tool," he said.