Hidden introns and repeated gains shape the heavily interrupted tRNA genes of Asgard archaea
Year: 2026
Type: preprint
Abstract: Transfer RNA (tRNA) genes are usually compact, yet in some Asgard archaea—the closest known prokaryotic relatives of eukaryotes—most are interrupted by introns. Studying this has been hampered by circularity: archaeal introns are catalogued by software that finds them partly through the splicing motif later used to explain where they persist. We built a motif-free census that detects intervening sequences by comparing each tRNA gene with compact relatives, and applied it to 90 Asgard genomes and 18 reference archaea. The standard annotator, tRNAscan-SE, had missed 30% of intervening sequences in the three focal families, mostly in the D-arm, but almost none in reference archaea; six independent checks supported the census. After correction, focal tRNA genes carry 2.2 intervening sequences on average, and in 30 genomes chosen before analysis the two focal families ranked first and second of 21 (exact P = 1 / 210). A reconstruction based on the census rather than on annotated introns placed 129 high-confidence gains at 65 loci on 37 branches of the species tree, 71 of them terminal; the two intron-rich classes are not sister lineages and carry mostly different intron positions, consistent with independent accumulation. Intervening sequences sat at canonical bulge–helix–bulge splicing motifs six to ten times more often than the average of three matched null models, including those the annotator missed and the reconstructed gains. Tests of copying between tRNA genes, mobile elements and reverse transcription found no general route of intron arrival. Asgard tRNA introns were thus gained repeatedly and sit preferentially at sites predicted to allow splicing; how they arrive remains open. Significance statement Every cell needs transfer RNAs to make proteins. In some Asgard archaea, the microbes most closely related to the ancestors of animals, plants and fungi, almost every tRNA gene is broken up by introns, which must be cut out before a tRNA can work. We show that the standard software for finding these introns misses nearly a third of the candidate intron sequences, and we provide a simple, motif-free way to find them. With a corrected catalogue, the evidence shows that introns were inserted many separate times, apparently independently in two lineages, and that they sit preferentially where the RNA can fold into the shape that the cell’s splicing enzyme is predicted to recognise. The study also shows how a detection tool can bias evolutionary conclusions about the very feature it detects, and how to test for that bias directly.
Institution: Visva-Bharati University
Language: English
Subfield: Molecular Biology
Domain: Life Sciences
Open Access status: green
OpenAlex ID: W7222172806