NEW YORK (GenomeWeb) – Researchers at the University of Colorado School of Medicine have developed a next-generation sequencing method to map the positions of modified bases such as uracil and pyrimidine dimers.
The method, called Excision-seq, combines in vitro nucleobase excision with next-generation sequencing on the Illumina platform and was published in Genome Research earlier this month.
Senior author Jay Hesselberth told In Sequence that his lab wanted to develop a method to study modified bases, but wanted an approach that was generalizable to different types of modifications, as opposed to one that targeted a specific type of modification. For instance, there are a number of sequencing-based methods specific to methylated cytosines and there are PCR-based methods to map nucleobase modifications, but Hesselberth wanted a method that could be applied to any type of modification and that was also scalable.
In the study, Hesselberth's team first demonstrated that the method could be used to map uracil in Escherichia coli and yeast. First, the researchers digested E. coli DNA with hypomorphic dUTPase, uracil DNA glycosylase, and T4 Endonuclease IV to remove the uracil and abasic sites, leaving a single-base gap. That created DNA fragments that could then be prepared for next-gen sequencing.
Testing on E. coli, the team found that they could reliably map uracil in the genome with a false positive rate of about 2 percent.
Next, they looked at a yeast strain known to have high levels of uracil. Somewhat surprisingly, they found a "huge variation in uracil content," Hesselberth said — a phenomenon the team is now following up on in a separate project. The finding could have implications for cancer therapy, he added. "A lot of cancer drugs change uracil content in DNA to promote cell death." But, uracil variation may work against the drugs, so understanding the variation could help design better drugs, he added.
Next, the team tested the method's ability to map a different type of modification — UV-induced pyrimidine dimers. In this case, the team used UVDE to cleave two types of pyrimidine dimers caused by UV light — cyclobutane pyrimidine dimers (CPDs) and 6-4 photoproducts. Because pyrimidine dimers are not compatible with subsequent PCR and adapator ligation techniques, the team first repaired the 5' pyrimidine dimers to "mono" pyrimidines with a photolyase enzyme. Then the team prepared the subsequent fragments for sequencing and aligned them to the yeast genome to determine the dipyrimidine locations.
More than 85 percent of the aligned sequences came from genomic positions with pyrimidine dimers, validating the method. In total, the team identified around 1.2 million sites of CPD formation and 107,490 sites of 6-4 photoproduct formation.
"Now, we're trying to set up experiments to look more carefully at those sites," Hesselberth said. "For example, there is some evidence that where you get a dimer is dictated by nucleosome position or methylation state, so we think the method could be useful for teasing out that stuff."
In addition, he said he is collaborating with a group at Johns Hopkins to use the method to look at DNA uracilization in relationship to HIV. There is an interesting phenomenon, Hesselberth said, in which HIV tends to become latent in immune cells with high levels of dUTP, and some evidence suggests that uracilization affects how HIV goes into the cells. In the collaboration, the groups plan to use Excision-seq to "understand how the virus gets uracilated at different stages during the infection process."
Hesselberth said he will also use the method to study methyl cytosine oxidation. "In principle, you can use any enzyme with Excision-seq" to look at any different type of base modification, he said.
Hesselberth acknowledged that, like Excision-seq, the Pacific Biosciences' RS II system can also detect base modifications and is not limited to one type of modification.
A feature of PacBio SMRT sequencing technology enables base modifications to be detected through the system's kinetics. Different modifications elicit a detectable pause in base incorporation. Currently, PacBio's software detects adenine and cytosine methylation, and reports out methyltransferase recognition motifs in bacterial genomes. Already, some researchers have used PacBio to study how differences in epigenetic modifications may alter pathogenicity of some bacteria.
However, Hesselberth said that for his purposes, he wanted a method that would have a higher throughput than the current PacBio system.