Hidden introns and repeated gains shape the heavily interrupted tRNA genes of Asgard archaea

Transfer RNA (tRNA) genes in certain Asgard archaea, considered the closest prokaryotic kin to eukaryotes, are frequently interrupted by introns. This phenomenon has been challenging to study due to circular reasoning: intron detection relies on software that finds them through splicing motifs, which are later used to explain their persistence.

To overcome this, researchers constructed a motif-free census that identifies intervening sequences by comparing each tRNA gene to compact relatives across 90 Asgard genomes and 18 reference archaea.

The standard annotator, tRNAscan-SE, missed 30% of these intervening sequences, predominantly in the D-arm, while the census detected almost none in reference archaea. Independent checks corroborated the census findings. Consequently, average Asgard tRNA genes now harbor 2.2 introns per gene, and two focal families emerged as the most intron-rich among 21 (P = 1/210). Reconstructing the species tree using the census instead of annotated introns identified 129 high-confidence gains at 65 loci across 37 branches.

Notably, these two intron-rich classes are not sister lineages and exhibit distinct intron positions, supporting independent accumulation. Introns predominantly appeared at canonical bulge-helix-bulge splicing motifs six to ten times more frequently than average, even in cases missed by the annotator and reconstructed gains. Investigations into copying mechanisms from tRNA genes, mobile elements, and reverse transcription revealed no consistent pathways for intron arrival.

Thus, Asgard tRNA introns were repeatedly gained and positioned favorably for splicing, but their origin remains uncertain.