Characterization of two bifunctional Arabdopsis thaliana genes coding for mitochondrial and cytosolic forms of valyl-tRNA synthetase and threonyl-tRNA synthetase by alternative use of two in-frame AUGs

We characterized two Arabidopsis thaliana cDNAs coding for class I valyl-tRNA synthetase and class II threonyl-tRNA synthetase. The proteins display characteristics of cytosolic enzymes, yet possess an N-terminal extension relative to their prokaryotic homologs. The proximal part of the N-terminal extension is a mitochondrial-targeting signal. Through transient expression of GFP fusions in tobacco cells, we demonstrated that both genes encode the cytosolic and mitochondrial forms of the enzymes by alternative use of two in-frame initiation codons. A long, mitochondrial form of the enzyme is translated from a first initiation codon at reduced levels because of a poor sequence context and a shorter, cytosolic form is translated from a second in-phase AUG, which is in a better context for translation initiation. Primer extension experiments revealed several transcript ends mapping upstream of the first AUG and between the two AUGs. Distal to the mitochondrial transit peptide both valyl-tRNA synthetase and threonyl tRNA synthetase possess an NH2-appended domain compared with their prokaryotic counterparts. This domain's amphiphilic helix is conserved between yeast and A. thaliana valyl-tRNA synthetase, suggesting an important role in translation. Based on the high structural similarities between yeast and A. thaliana valyl-tRNA synthetase, we propose that the acquisition of bifunctionality of valyl-tRNA synthetase predates the divergence of these two organisms.