Different pattern ofcodonrecognition bymammalian mitochondrial tRNAs

Analysis ofanalmost cornplete mammalian mitochondrial DNAsequence hasidentifiemd23 possible tRNA genes andwespeculate here that these aresufficient totranslate all thecodons ofthemitochondrial genetic code. This number ismuchsmaller than theminimum of31required bythewobble hypothesis. Foreachoftheeight genetic codeboxes with four codons foroneaminoacid wefind asingle specific tRNAgene with Tinthefirst (wobble) position oftheanticodon. Wesuggest that these tRNAswith Uinthewobble position canrecognize all four codons inthese genetic codeboxes either bya"twoout ofthree" base interaction orbyUN wobble. Mitochondrial DNAcodes foranumber oftRNAsaswell as ribosomal RNAsandseveral respiratory proteins. Nineteen tRNAshavebeenmapped (byhybridization) onthehuman mitochondrial genome (1, 2)andmitochondrial tRNAs specific for all amino acids except L-proline, glutamine, histidine, and asparagine have beenidentified bycharging experiments (1). Similar studies onratliver mitochondrial tRNAshave identified all species except those forglutamine andhistidine (3). As- suming that all mammalian mitochondria codeforthesame complement oftRNAs, these studies bring thenumber of mammalian mitochondrial tRNAs sofar identified to21,which includes twoserine tRNAs, twoleucine tRNAs, atRNAmet, and atRNAMet. Studies onpurified ratliver mitochondria and cytoplasmic tRNAshavefailed toshowanyimportation of cytoplasmic tRNAsinto themitochondrion (3). Themito- chondrial tRNAshaddifferent chromatographic properties on RPC-5columns andpolyacrylamide gelelectrophoresis from their cytoplasmic counterparts, andnomitochondrial tRNAs hybridized tonuclear DNA. Thesmall number ofmammalian mitochondrial tRNAsand thelack ofanyimported cytoplasmic tRNAsraises theproblem ofhowthese tRNAsrecognize all possible codons inmito- chondrial protein synthesis. Onthebasis ofthewobble hy- pothesis (4), aminimumof32tRNAsareneeded toread all possible codons ofthe"universal" genetic code. Recent studies have established that humanandbovine mitochondria have a different genetic code inthat UGAistryptophan andnotter- mination andAUAismethionine andnotisoleucine (ref. 5; I.G.Y., unpublished results). Thus, aminimumof31tRNAs would beneeded. However, recently anumber ofinvitro experiments have indicated that thewobble hypothesis might notapply andthat insomecases thethird (wobble) position ofthecodon hasno discriminatory function. Mitra etal.(6) studied theincorpo- ration ofvaline inaninvitro protein-synthesizing