Glycyl tRNA synthetase (GlyRS) provides a unique case among class II aminoacyl tRNA synthetases, with two clearly widespread types of enzymes: a dimeric (?<
sub>
2<
/sub>
) species present in some bacteria, archaea, and eukaryotes
and a heterotetrameric form (?<
sub>
2<
/sub>
?<
sub>
2<
/sub>
) present in most bacteria. Although the differences between both types of GlyRS at the anticodon binding domain level are evident, the extent and implications of the variations in the catalytic domain have not been described, and it is unclear whether the mechanism of amino acid recognition is also dissimilar. Here, we show that the ?-subunit of the ?<
sub>
2<
/sub>
?<
sub>
2<
/sub>
GlyRS from the bacterium Aquifex aeolicus is able to perform the first step of the aminoacylation reaction, which involves the activation of the amino acid with ATP. The crystal structure of the ?-subunit in the complex with an analog of glycyl adenylate at 2.8 � resolution presents a conformational arrangement that properly positions the cognate amino acid. This work shows that glycine is recognized by a subset of different residues in the two types of GlyRS. Furthermore, a structural and sequence analysis of class II catalytic domains shows that bacterial GlyRS is closely related to alanyl tRNA synthetase, which led us to define a new subclassification of these ancient enzymes and to propose an evolutionary path of ?<
sub>
2<
/sub>
?<
sub>
2<
/sub>
GlyRS, convergent with ?<
sub>
2<
/sub>
GlyRS and divergent from AlaRS, thus providing a possible explanation for the puzzling existence of two proteins sharing the same fold and function but not a common ancestor.