Mapping RNA modifications in the human brain with full-length transcript sequencing

A single gene can produce multiple RNA isoforms through complex gene regulation mechanisms, such as alternative splicing and post-transcriptional modifications. This creates a large transcriptional diversity by generating multiple isoforms, as exhibited in the human brain, facilitating the regulation of a wide range of neurological functions, such as synaptogenesis and axon guidance¹.

N6-methyladenosine (m⁶A) is a post-transcriptional modification that is most abundant in the human brain¹ and is defined by the addition of methyl groups to adenine residues². This modification is exhibited at the highest levels in the human brain and its abundance increases during developmental stages through to adulthood¹. It is known to control the stability of mRNA and regulate neuronal activity², as well as playing a critical role in brain development, memory, and learning¹.

Despite the abundance of m⁶A and its known critical role in brain development, little is known about the modification at the isoform level, nor how it is influenced by isoform structure and polyadenylation. Profiling this modification may shed light on normal brain development and the mechanisms behind diseases — including dysregulation of RNA modifications, which have been implicated in many neurodegenerative and psychiatric disorders, such as schizophrenia and Alzheimer's disease².

Traditional methods rely on short-read data to detect m⁶A modifications. Short-read sequencing is typically used with traditional sample preparation methods, such as immunoprecipitation-based techniques, which only provide information at the gene level, leaving the original RNA isoform containing m⁶A modifications unidentified¹. Prior to sequencing, the expensive and laborious sample preparation methods only enrich short fragments of modified RNA before cDNA conversion, dissociating the modification from the native context³, and can introduce sequencing bias, potentially obscuring important biology. Furthermore, short-read sequencing techniques do not provide nucleotide-level resolution of modifications³ and can exhibit high rates of multimapping, where short reads cannot be assigned to a single transcript, hindering accurate computational assembly.

To ‘provide an isoform-level transcriptome-wide map of m⁶A modification sites’, Gleeson et al. utilised direct RNA nanopore sequencing to generate long reads, spanning full-length transcripts from functionally distinct regions of the brain¹. PCR-free nanopore sequencing allows the direct sequencing of native RNA without requiring conversion to cDNA. This enables quantification and comprehensive transcriptome characterisation at the isoform level, supporting the generation of novel insights into the m⁶A modification and its role in brain function and disease.

‘Long-read direct RNA sequencing from Oxford Nanopore Technologies addresses many of these limitations by providing single-nucleotide isoform-level resolution of m⁶A modifications’

The researchers sequenced tissue research samples from three functionally distinct regions of the brain: the prefrontal cortex, caudate nucleus, and cerebellum. The RNA was extracted from post-mortem research samples that had no diagnosis, nor physiological evidence of neurological or neuropsychiatric disorders. The team utilised the Direct RNA Sequencing Kit to prepare the isolated poly-A RNA for sequencing on a PromethION device.

The sequencing data mapped m⁶A modifications and polyadenylation across the transcriptome, revealing widespread differences of m⁶A modifications in isoforms between brain regions. From ten samples, >52 million high-quality reads were generated. With the long nanopore reads, the researchers identified the expression of >22,000 genes and >62,000 isoforms across the brain regions.

Markedly, the long nanopore reads enabled the ‘identification of m⁶A modification sites at the isoform level with single nucleotide resolution’, allowing the team to determine the exact transcriptomic position of a modification. Widespread changes in isoform expression, m⁶A profiles, and poly-A tail lengths were also consistently identified between gene isoforms and brain regions. Particularly, expression patterns of m⁶A machinery — such as m⁶A writers and erasers — as well as isoform architecture, reflected the m⁶A modification levels in different brain regions. For example, the highest proportion of m⁶A modifications were found in the cerebellum region — an area with increased m⁶A writers — whereas the caudate nucleus exhibited the lowest modification rates at m⁶A sites, likely due to an increase of m⁶A erasers.

The researchers further demonstrated nanopore technology can sequence poly-A tails, which are critical in post-transcriptional regulation to stabilise mRNA and promote translation. Utilising nanopore sequencing, the team measured and identified tail length differences between brain regions and observed that poly-A tail length was associated with isoform expression and m⁶A modification levels; genes that encoded for multiple isoforms with distal polyadenylation sites had increased m⁶A modification rates and longer poly-A tails. Furthermore, the number of m⁶A sites on an isoform was moderately correlated with poly-A tail length, further highlighting the ‘complexity of gene regulation and [suggesting] that the interplay of polyadenylation and m⁶A modification patterns contribute to regulating gene expression in the brain’.

‘Oxford Nanopore direct RNA sequencing can quantify isoform expression, modifications and poly-A tail lengths, enabling simultaneous investigation of the transcriptome and epitranscriptome’

The nanopore data demonstrates that there are multiple RNA regulatory mechanisms working in conjunction with each other, and the integration of nanopore sequencing opens up opportunities by ‘providing single-nucleotide isoform-level resolution of m⁶A modifications’. Furthermore, the nanopore data suggests that traditional sequencing methods have possibly ‘masked many isoform-specific regulation events of m⁶A deposition’ as they only provide modification information at the gene level, whereas the nanopore data provides isoform-level resolution.

In conclusion, Gleeson et al. revealed novel isoform-level and brain region-specific patterning of m⁶A modifications and polyadenylation with direct RNA nanopore sequencing. Their research further highlighted how the application of nanopore sequencing will support a ‘move towards understanding the functional implications of m⁶A modifications’, and their role in neurological development and disease.