Back

Conservation of function without conservation of amino acid sequence in intrinsically disordered transcriptional activation domains

LeBlanc, C. J.; Stefani, J.; Soriano, M.; Lam, A.; Zintel, M. A.; Kotha, S. R.; Chase, E.; Pimentel-Solorio, G.; Vunnum, A.; Flug, K. L.; Fultineer, A. T.; Hummel, N.; Staller, M. V.

2024-12-05 systems biology
10.1101/2024.12.03.626510 bioRxiv
Show abstract

Protein function is canonically believed to be more conserved than amino acid sequence, but this idea is only well supported in folded domains, where highly diverged sequences can fold into equivalent 3D structures with identical function. Intrinsically disordered protein regions (IDRs) often experience rapid amino acid sequence divergence, but because they do not fold into stable 3D structures, it remains unknown when and how function is conserved. As a model system for studying the evolution of IDRs, we examined transcriptional activation domains, the regions of transcription factors that bind to coactivator complexes. We systematically identified activation domains on 502 homologs of the transcriptional activator Gcn4 spanning 600 MY of fungal evolution in the Ascomycota. We find that the central activation domain shows strong conservation of function without conservation of sequence. We identify the molecular mechanism for this conservation of function without conservation of sequence: evolutionary turnover (gain and loss) of acidic and aromatic residues that are important for function. We further see turnover of complete N-terminal activation domains. This turnover at two length scales confounds multiple sequence alignments, explaining why traditional comparative genomics cannot detect functional conservation of activation domains. Evolutionary turnover of key residues is likely a general mechanism for conservation of function without conservation of sequence in IDRs.

Matching journals

The top 5 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.