Microexons: highly conserved protein microsurgery during neuronal maturation

Against common intuition, tiny microexons as short as 3 nucleotides can and do exist within our genes. Surprisingly, not only can they be recognized by the cellular machinery and become part of proteins, but they were also often found to be only (or mainly) expressed in neurons. Microexons are switched on during late neuronal differentiation, and impact surfaces of proteins that are crucial for neuronal maturation and synaptic function. Through this protein surface 'microsurgery', microexons can modulate how proteins interact with each other, and their misregulation is associated with some individuals with autism spectrum disorder.

HFSP Long-Term Fellow Manuel Irimia and colleagues
authored on Fri, 06 February 2015

Since the discovery, almost four decades ago, that genes are encoded in pieces, many researchers have investigated the mechanisms by which the ‘spacer’ sequences (introns) are removed from the transcripts while the ‘meaningful’ sequences (exons) are spliced together to form the mature mRNAs. Following this discovery, one of the most exciting findings was that, by selectively removing or inserting specific exons, this mechanism allows for the production of multiple mRNA and protein isoforms from each gene, sometimes with dramatically different functions, in a process known as alternative splicing. Understanding the biological roles of alternative exons is currently a major topic of research.

Figure: Artwork by Alexander Weiss

Despite a few known exceptions, a general assumption in the field has been that exons need to be relatively large (averaging ~145 nucleotides in humans), so that the cellular machinery can properly recognize them within the surrounding sea of intronic sequence (generally thousands of nucleotides). However, in this study we identified hundreds of tiny microexons in our genome, between 3-27 nucleotides in length. Surprisingly, microexons can not only be recognized by the spliceosome, but most of them show a remarkably tight neural-specific regulation. Dozens of microexons are spliced-in only in the transcripts produced in the neurons, and they present switch-like regulation of their inclusion during late neuronal differentiation. Importantly, we identified the neuronal- and vertebrate-specific splicing factor SRRM4/nSR100 as a major direct regulator of these switches. Strikingly, they are also the most conserved type of alternative splicing observed so far, and we identified many neural-specific microexons conserved from human to shark, spanning 450 million years of vertebrate evolution.


What are the functions of these neuronal microexons? We observed that nearly all of them were multiples of three nucleotides and thus maintained protein reading frames, producing specific protein isoforms in neurons that diverge from their non-neuronal counterparts by as little as 1 or 2 amino acids. We thus asked how these tiny differences could affect protein function. We found that microexons are often located in protein surfaces associated with binding domains, where they may subtly modulate domain functions without disrupting their structure. We thus hypothesized that microexons could impact how proteins interact with each other. As proof of concept, we experimentally demonstrated that microexons do in fact affect interactions in proteins that are important for synaptogenesis and learning such as APBB1 and the AP1 complex.

Finally, we reasoned that if microexons are so evolutionarily conserved, have striking switch-like regulation during neuronal maturation and can impact protein interactions, they should play important biological functions and thus their misregulation could be associated with neural disorders, particularly those showing synaptogenic and maturation defects. By comparing deep RNA sequencing of brains of individuals with autism and matched controls, we found that a group of autistic individuals show significant misregulation of a subset of microexons.  Remarkably, some of these microexons are in genes that have been traditionally associated with this disorder, opening an entirely new avenue of research to understanding the molecular bases of autism.


A highly conserved program of neuronal microexons is misregulated in autistic brains.Irimia, M., Weatheritt, R.J., Ellis, J., Parikshak, N.N., Gonatopoulos-Pournatzis, T., Babor, M., Quesnel-Vallières, M., Tapial, J., Raj, B., O’Hanlon, D., Barrios-Rodiles, M., Sternberg, M.J.E., Cordes, S.P., Roth, F.P., Wrana, J.L., Geschwind, D.H., Blencowe, B.B. 2014. Cell, 159:1511-23.

Pubmed link

Microexons go big (Cell)

Microexons on the brain (Nature Genetics)

Different ways to splice the cake (Nature Reviews Neuroscience)

Microexons tiny but mightly (EMBO Press)

SFARI website