The Escherichia coli su+3 tyrosine tRNA was shown recently to be a leucine-specific tRNA in Saccharomyces cerevisiae. This finding raises the possibility that some determinants for tRNA identity in E. coli may be different in S. cerevisiae. To investigate whether the fungal system is sensitive to the major determinant for alanine acceptance in E. coli, a single G3. U70 base pair was introduced into the acceptor helix of the su+3 tyrosine tRNA. This substitution converts the identity of the E. coli suppressor in S. cerevisiae from leucine to alanine. Thus, as in E. coli, G3-U70 is a strong determinant for alanine acceptance that can dominate over other features in a tRNA that might be recognized by alternative charging enzymes.In principle, the determinants for the identity of a tRNA can differ from organism to organism without compromising the conservation of the genetic code. The only requirement is that the anticodon triplet of each tRNA must specify the same amino acid from one organism to another. In the case of Escherichia coli alanine tRNA, a single G3 U70 base pair is a major determinant for identity (9, 13). Substitution of this base pair with G C, A. U, or U G abolishes aminoacylation with alanine, and transfer of G3 U70 into E. coli tRNACYS, tRNAPhe, and tRNATYr can confer total or partial alanine acceptance on each (13,14,21,25). Cytoplasmic Bombyx mori and human alanine tRNAs each encode a G3 U70 base pair, and the aminoacylation of these tRNAs by their homologous enzymes is abolished when that base pair is changed to G C or A. U. Moreover, expression of these eukaryotic tRNAs in E. coli results in charging with alanine and no other detectable amino acid (15). Thus, in the evolution of these alanine tRNAs, the G3 -U70 base pair has been conserved. Furthermore, nucleotide sequence differences in other parts of the B. mori and human cytoplasmic alanine tRNAs do not encode major determinants for any other E. coli aminoacyl-tRNA synthetase.In contrast to the results obtained with alanine tRNAs, the identity of the E. coli su+3 tyrosine tRNA is changed in the yeast Saccharomyces cerevisiae. This amber suppressor inserts only tyrosine in E. coli, but in S. cerevisiae only leucine is inserted at the position of an amber codon (7). This change in specificity may correlate with a switch in the type class of cytoplasmic tyrosine tRNAs in comparisons of E. coli with S. cerevisiae. In E. coli, tyrosine, leucine, and serine tRNAs are type II tRNAs, which have three base pairs in the dihydrouridine stem and 13 to 22 nucleotides in the variable loop (also known as the extra arm). The remaining E. coli tRNAs are type I tRNAs which typically have four base pairs in the dihydrouridine stem and five nucleotides in the variable loop. While the cytoplasmic leucine and serine tRNAs are also type II structures in S. cerevisiae, the tyrosine tRNA is a type I molecule. Thus, the switch in identity of the su+3 tyrosine tRNA in S. cerevisiae may mean that, in the evolution of some tRNAs, molecules in the The mode of recognitio...