NEW YORK – Researchers investigating methylation patterns in RNA have a new suite of high-resolution, sequencing-based tools that don't rely on expensive antibodies.
"There seems to be an ever-growing list" of technologies, which are driving this field, said Kate Meyer, a professor at Duke University, who was first author on the first paper to introduce antibody-based sequencing approaches for detecting N6-methyladenosine (m6A) modifications in RNAs in 2012. Now, she's made another leap in technology with Dart-seq, which stands for "deamination adjacent to RNA modification targets sequencing."
Not only is it cheaper, the method also requires much less sample material: as little as 10 nanograms, Meyer said, compared to 10 micrograms for antibody-based approaches. Meyer published a paper describing Dart-seq last month in Nature Methods.
Dart-seq is one of several high-resolution approaches to discovering these RNA modifications. "It's definitely an improvement on almost every other method that gives you a region of methylation," said Chris Mason, a researcher at Weill Cornell Medicine, who has worked on several approaches to detect m6A modifications in RNA. Mason was not involved in Dart-seq, which Meyer said she developed alone, but they co-authored the paper that introduced the first antibody-based sequencing approach. He was also involved in a new nanopore sequencing-based approach to detecting m6A modifications, described last month in Nature Communications.
In July, researchers led by Guan-Zheng Luo of Sun Yat-sen University in Guangzhou, China, and Chuan He of the University of Chicago, revealed another method for mapping m6A without antibodies, called m6A-sensitive RNA-Endoribonuclease-Facilitated sequencing, or m6A-REF-seq. They published their paper in Science Advances.
And m6A is just one of more than 100 epigenetic modifications that can be found on RNAs. Researchers are developing sequencing-based methods for those, too.
"We have not one, not two, but almost a dozen different methods that work on different platforms," Mason said. "The capacity to see this biology has never been larger."
The m6A modification is a prime target because it's one of the more prevalent modifications seen in eukaryotic mRNA. But the field is so new, "there's a lot that's left to be worked out," Meyer said.
Just seven years ago, she, Mason, and senior author Samie Jaffrey, among others, published on methylated RNA immunoprecipitation sequencing (MeRIP-Seq) which used antibodies to pull out RNAs with the m6A modifications. The method opened studies on this epigenetic mark and joined methods like bisulfite sequencing.
But antibodies are expensive and MeRIP-Seq can't provide single-base resolution. Moreover, it requires a lot of RNA. "It's been a great technique, but by and large the question I heard the most is, 'Do we really need this much RNA? Because it's not feasible for our study,'" Meyer said. Now a tenure-track junior professor at Duke, Meyer has been working on Dart-seq — her "side project" — for about a year and a half.
The method uses engineered proteins to flip bases near m6A sites. The key is fusing a cytosine-to-uracil base flipping protein, APOBEC1, to the YTH protein domain, which is present in five human proteins, all of which bind m6A, Meyer said. The recent study read out those flipped bases using Illumina sequencing, but she said that any platform would work.
"The advantage is you get single-base resolution," Mason, who reviewed the Dart-seq paper, said. "But it also showed some degree of quantification at the same time." Meyer added that Dart-seq can distinguish between m6A and N6,2′-O-dimethyladenosine (m6Am) while some antibodies cannot due to cross reactivity.
Dart-seq does require some advance scouting. Meyer said it's necessary to look at natural C-to-U transitions that might be present in the sample genome and that SNPs must be accounted for. Another control study was to look at all the base flips APOBEC1 would induce without being fused to the m6A binding domain.
The competing m6a-REF-seq method also uses proteins to help identify modification sites. Here, the researchers used m6a-sensitive endoribonucleases to cleave the RNAs and then ligate them to sequencing adapters. They also used Illumina sequencing. In their paper, the researchers wrote that their results led to quantifiable, single-base mapping of m6A, which "showed that two adjacent m6A sites were statistically prone to aggregate within 200-nt regions… supporting the hypothesis that m6A functions as clusters of modification."
And last month, researchers led by Eva Maria Novoa, of the Barcelona Institute of Science and Technology, and including Mason, published a paper in Nature Communications showing that other sequencing platforms could help identify RNA modifications. Their paper used Oxford Nanopore Technologies' platform to identify m6A.
"We hypothesized that the current intensity changes caused by the presence of RNA modifications may lead to increased 'errors' and decreased qualities from the output of base-calling algorithms that do not model base modifications," the researchers wrote. "Indeed, here we find that base-calling 'errors' can accurately identify m6A RNA modifications in native RNA sequences, and propose a novel algorithm, EpiNano, which can be used to identify m6A RNA modifications from RNA reads with an overall accuracy of [approximately] 90 percent."
The results, they noted, provide proof of concept for the use of base-called features to identify RNA modifications in general using direct RNA sequencing.
Applications and commercialization
"In the paper, we're just introducing the technique," Meyer said of Dart-seq. New technology has been driving the field, she said, namely "the ability to uncover and measure abundance" of modifications and "how and if they're changing." Because the modifications are found in thousands of messenger RNAs, as well as long non-coding RNAs, she expects them to play a role in "every aspect of the RNA lifecycle."
These sequencing-based methods may disrupt the market for m6A antibodies, which are sold by several firms, including Thermo Fisher Scientific.
For academic researchers, Meyer's plasmid is available on non-profit plasmid repository Addgene, but she and Duke are exploring commercialization. She filed for a patent on the method and while she said her priority is "getting the technology out there to as many hands as possible," if there's a financial opportunity, she's open to that, too.
"Duke is in conversations with a number of companies to license it," said Christy Ferguson, associate director of licensing at Duke; however, she declined to say which ones. "Initially, it will probably be used as a research tool; eventually our hope is that it will be useful for clinical applications. Our licensing conversations are still in an early stage, but there is a lot of commercial interest."
A spinout is another possibility, but Meyer notes she's "still an assistant professor, trying to get tenure."
"My lab is only three years old," she said. "I'd have to weigh how much time it would take away, but at this point I'm entirely open to it."