Probably the only RNA molecule to be imported into yeast mitochondria is tRNA1Lys. The tRNA is not aminoacylated in mitochondria and its function remains unclear. I have noted that after partial unwinding the tRNA1Lys (more of less in the position, where reverse transcriptase unwinds primer tRNAs) it is possible to fit the donor regions of seven different mitochondrial mosaic genes introns to the unwound part of tRNA. The exact splice site is in all cases found according to the following rule: base pair, cross, locate first m2G X U, cut, base pair. Possible exception is the 5th intron of cob-box where the boundary indicated is not the one accepted in literature, nevertheless the boundary suggested here seems possible. The complementarity suggested permits accurate prediction of 4 (5 if b15 counts) XuncanonicalX donor boundaries. The four remaining introns are fitted by the author into looped structures held by unwound tRNA1Lys. The two groups of introns suggested coincide exactly with the Dujonian classes I and II, his classification being based on inner clusters of homology rather than intron/exon boundaries used here. A quite impressive fit was found also between tRNA1Lys and box 9, a sequence needed for splicing of a 1st class intron b14 RNA. The sequences playing probably a crucial role in splicing of 2nd class introns: GGGGG and GGCUC are both complementary to a 3'-CCCCC--GAG-5'-sequence in the GT psi C loop+arm of tRNA1Lys. Used or not in splicing tRNA1Lys curiously enough looks like a compendium of nucleotide sequences needed to handle the mitochondrial mosaic genes RNAs.(ABSTRACT TRUNCATED AT 250 WORDS)