Decoding the function(s) of the m6A modifications on viral RNAs - Daniel Depledge
Daniel P. Depledge, Assistant Professor at the New York University (NYU) School of Medicine, opened the RNA modifications breakout by discussing his team’s latest research into the functions of N6-methyladenosine (m6A) modifications on viral RNAs. Daniel explained how m6A is one of the most prevalent modifications of messenger RNA (mRNA) in eukaryotic cells. It has also been shown to play important proviral or antiviral roles in the life cycles of a growing number of viruses. Due to the extremely compact nature of dsDNA viral genomes and the presence of many overlapping RNAs, the identification of m6A modifications at the isoform level using traditional short-read sequencing technologies is a considerable challenge.
Building on earlier work, where the team at the NYU School of Medicine demonstrated that direct RNA sequencing using nanopore technology enables the identification of splicing patterns, transcription initiation, and polyadenylation in cells infected with dsDNA viruses (including herpesviruses and adenoviruses), the team have now applied the benefits of direct RNA sequencing to characterise m6A modifications at nucleotide resolution. Critically, they were also able to distinguish between overlapping RNA isoforms that are differentially m6A modified.
The team utilised adenovirus as a model for RNA synthesis. Adenovirus infection is divided into two phases, ‘early’ and ‘late’. During the ‘early’ phase cellular replication factors are induced that initiate viral genome replication, host antiviral responses are subverted, and host cellular processes are rewired towards late protein production. In the ‘late’ phase, late promoters are activated, resulting in the production of late proteins, alternative splicing, and the assembly of new virions. Dan stated that it is imperative that you have a really good viral genome annotation before looking for modifications. To this end, the first part of the project involved using direct RNA sequencing to update their annotation of adenovirus, which increased the number of transcriptional units from approximately 20 to over 75.
Dan showed western blotting and cell staining data indicating that during an adenovirus infection of the cell the m6A machinery is recruited to sites of viral RNA synthesis.
Using the short-read sequencing meRIP approach, which utilises m6A-specific antibodies to pull-down regions enriched for m6A, the team were able to identify modifications within a specific region of RNA but not at nucleotide resolution. Dan stated that the meRIP approach is “nice, but not perfect” as it is sensitive, but not specific. As such, it is not compatible with transcript-level analysis, which is key to understanding the function of m6A modifications.
To overcome these challenges the team used direct, long-read nanopore sequencing to analyse two adenovirus-infected A549 cell lines (a parent cell line and a METTL3 knockout line). The resultant sequence was aligned to the genome (and later transcriptome), and every nucleotide position for each read across both cell lines was examined to show the numbers of A, G, C, U, and N (indels) nucleotides. Statistical analysis of this data allowed the identification of positions between the two cell lines at where there is a divergence. Reassuringly when looking at an adenosine modification, the vast majority of the sites identified using this approach were indeed As. Next, the team examined the distance between each modification, finding, in many cases, it to be just 1-3 nucleotides. They then introduced a masking strategy that assigned any modification within a 5-nucleotide radius to the nucleotide with the highest G-test value. Examining this data revealed that the majority of modifications were identified within AC motifs, which are the minimal motifs within which m6A can occur. Overlaying this data to the genome map indicated specific positions at which m6A is believed to exist. Recapitulating this statistical analysis, but using alignment against the transcriptome, identified an even greater number of putative modification sites. Excitingly, Dan also showed how this methodology allowed the mapping of modifications to specific alternative transcripts, a result that he described as ‘mind blowing’.
Closing the presentation, Dan went on to describe how initial functional studies using this methodology suggest that m6A appears to affect splicing efficiency of late viral RNA’s.
Check out Dan's latest pre-print from 5th Dec, on m6A detection, in our Resource Centre: https://nanoporetech.com/resource-centre/direct-rna-sequencing-reveals-m6a-modifications-adenovirus-rna-are-necessary