Photoreactive derivatives of yeast tRNA Phe containing 2-azidoadenosine at their 3P P termini were used to trace the movement of tRNA across the 50S subunit during its transit from the P site to the E site of the 70S ribosome. When bound to the P site of poly(U)-programmed ribosomes, deacylated tRNA Phe , Phe-tRNA Phe and N-acetyl-Phe-tRNA Phe probes labeled protein L27 and two main sites within domain V of the 23S RNA. In contrast, deacylated tRNA Phe bound to the E site in the presence of poly(U) labeled protein L33 and a single site within domain V of the 23S rRNA. In the absence of poly(U), the deacylated tRNA Phe probe also labeled protein L1. Crosslinking experiments with vacant 70S ribosomes revealed that deacylated tRNA enters the P site through the E site, progressively labeling proteins L1, L33 and, finally, L27. In the course of this process, tRNA passes through the intermediate P/E binding state. These findings suggest that the transit of tRNA from the P site to the E site involves the same interactions, but in reverse order. Moreover, our results indicate that the final release of deacylated tRNA from the ribosome is mediated by the F site, for which protein L1 serves as a marker. The results also show that the precise placement of the acceptor end of tRNA on the 50S subunit at the P and E sites is influenced in subtle ways both by the presence of aminoacyl or peptidyl moieties and, more surprisingly, by the environment of the anticodon on the 30S subunit. ß