Phenylalanyl-tRNA synthetase from the archaebacterium Methanosarcina barkeri activates a number of phenylalanine analogues (methionine, p-fluorophenylalanine, beta-phenylserine, beta-thien-2-ylalanine, 2-amino-4-methylhex-4-enoic acid and ochratoxin A) in the absence of tRNA, as demonstrated by Km and kcat of the ATP/PPi exchange reaction. Upon complexation with tRNA, AMP formation from the enzyme X tRNA complex in the presence of ATP, one of the above analogues or tyrosine, leucine, mimosine, N-benzyl-L- or N-benzyl-D-phenylalanine indicates activation of the analogues under conditions of aminoacylation. Natural noncognate amino acids are not transferred to tRNAPhe-C-C-A or tRNAPhe-C-C-A-(3'-NH2). This pretransfer proofreading mechanism, together with the comparatively low ratio of synthetic to successive hydrolytic steps, resembles the mechanism of liver enzymes of vertebrates. In contrast, eubacterial phenylalanyl-tRNA synthetases achieve the necessary fidelity by post-transfer proofreading, a corrective hydrolytic event after transfer to tRNAPhe. Diadenosine 5',5'''-P1,P4-tetraphosphate synthesis is shown to be a common feature for phenylalanyl-tRNA synthetases from all three lineages of descent. The immunological approach demonstrates that aminoacyl-tRNA synthetases do not belong to the group of enzymes in gene expression with high structural conservation.