The conserved positions of the eukaryotic cytoplasmic initiator tRNA have been suggested to be important for the initiation of protein synthesis. However, the role of these positions is not known. We describe in this report a functional analysis of the yeast initiator methionine tRNA (tRNAVet), using a novel in vivo assay system which is not dependent on suppressor tRNAs. Strains of Saccharomyces cerevisiae with null alleles of the four initiator methionine tRNA (IMT) genes were constructed. Consequently, growth of these strains was dependent on tRNAe et encoded from a plasmid-derived gene. We used these strains to investigate the significance of the conserved nucleosides of yeast tRNA,Met in vivo. Nucleotide substitutions corresponding to the nucleosides of the yeast elongator methionine tRNA (tRNAmet) have been made at all conserved positions to identify the positions that are important for tRNA,4et to function in the initiation process. Surprisingly, nucleoside changes in base pairs 3-70, 12-23, 31-39, and 29-41, as well as expanding loop I by inserting an A at position 17 (A17) had no effect on the tester strain. Nucleotide substitutions in positions 54 and 60 to cytidines and guanosines (C54, G54, C60, and G60) did not prevent cell growth. In contrast, the double mutation U/rT54C60 blocked cell growth, and changing the A-U base pair 1-72 to a G-C base pair was deleterious to the cell, although these tRNAs were synthesized and accepted methionine in vitro. From our data, we suggest that an A-U base pair in position 1-72 is important for tRNA,eet function, that the hypothetical requirement for adenosines at positions 54 and 60 is invalid, and that a U/rT at position 54 is an antideterminant distinguishing an elongator from an initiator tRNA in the initiation of translation.In the translational decoding of mRNA, codons are read by a unique set of tRNA species. Each isoacceptor is usually encoded by more than one copy of the gene. To study the function of a tRNA in protein synthesis, a common strategy is to generate tRNA mutations. To date, phenotypic changes in vivo have been scored either by different types of suppressor tRNAs (nonsense, missense, and frameshift, i.e., a tRNA with an already existing mutation) or by an in vivo system in which the rate of aminoacyl-tRNA selection is measured in competition with a natural tRNA'U frameshifter (14,19,39). The former strategy has some limitations. First, the anticodon is a major identity element for some tRNA synthetases (41, 49), including the methionyl-tRNA synthetase from Escherichia coli (50). Therefore, the translational efficiency of the suppressor tRNA cannot be distinguished from its capacity to be aminoacylated. Second, in the translational process, a stop codon is a signal for the release factors to stop translation. Thus, the translational efficiency of the suppressor tRNA will be masked, as the tRNA must compete with release factors for the stop codon. In this report, an in vivo system which overcomes these obstacles by being dependent on one single I...