The molecular events leading to the synthesis of mature tRNA are only now becoming amenable to experimental study. In bacterial and mammalian cells tRNA genes are transcribed into precursor tRNA. These molecules, when isolated, contain additional nucleotides at both ends (20) of the mature tRNA and lack most modified nucleosides. Presumably, specific nucleases ("trimming" enzymes) cut the precursor to proper tRNA size. The C-C-A nucleotide sequence of the amino acid acceptor end common to all tRNA's does not seem to be coded by tRNA genes (30), and may be added to the trimmed molecules by the tRNA-CMP-AMP-pyrophosphorylase (71). Modifications at the polynucleotide level of the heterocyclic bases or the sugar residues give rise to the modified nucleosides in tRNA. Although newly available substrates have allowed the detection of more of the enzymes involved in these reactions, there is still no knowledge about the sequence of modification or trimming events leading to the synthesis of active tRNA. Progress in these studies may not be easy because enzyme preparations free of nucleases or other tRNA modifying enzymes are required. The role of the modified nucleosides in the biological functions of tRNA is still unknown. Possibly pseudouridine is required for ribosome mediated protein synthesis; some other modified nucleosides in tRNA are not required for this reaction, but may enhance its rate. What might be the role of the large variety of modified nucleosides in tRNA? One is tempted to speculate that such nucleosides are important in other cellular processes in which tRNA is thought to participate such as virus infection, cell differentiation, and hormone action (2, 3). Mutants in a number of tRNA-modifying enzymes are needed in order to extend our knowledge of their purpose and of tRNA involvement in other biological processes. But unless tRNA-modifying enzymes specific for a particular tRNA species exist, no simple selection procedure can be devised. Possibly some of the regulatory mutants of amino acid biosynthesis may prove to affect tRNA-modifying enzymes (72). Transfer RNA's are macromolecules well suited for the study of nucleic acid-protein interactions. The tRNA molecules are structurally very similar, and they interact with a large number of enzymes or protein factors (2, 3). Each aminoacyl-tRNA synthetase, for instance, very precisely recognizes a set of cognate isoacceptor tRNA's (2, 73). The availability of the tRNA- modifying enzymes adds another dimension to the problem of the nature of specific recognition of tRNA by proteins. There are some tRNA-modifying enzymes, such as the uracil-tRNA methylase, which may recognize all tRNA species, while others, such as the isopentenyl-tRNA transferase, probably recognize only a selected set of tRNA molecules, even with different amino acid accepting capacities. With well-characterized RNA precursor and tRNA molecules we can hope to delineate those features of primary, secondary, and tertiary structure involved in the specific interactions of tRNA with these enzymes.