PS decarboxylases in Arabidopsis. A, Amino acid sequence comparison of Arabidopsis PSD1 (At4g16700), PSD2 (At5g57190), and PSD3 (At4g25970). Arabidopsis contains three genes with sequence similarities to PS decarboxylases. In contrast to the mitochondrial PS decarboxylase, PSD1, the PSD2 and PSD3 proteins contain an N-terminal extension of approximately 350 amino acids. A truncated form of PSD3 starting with the amino acid Ser-352 (indicated with an asterisk) was expressed in E. coli ( Fig. 2B). Identical amino acids are highlighted in black, and gaps are indicated with dashes. The conserved GSTV motif is marked with a box. B, Phylogenetic tree of PS decarboxylases. Amino acid sequences (the C-terminal 300 amino acids without N-terminal extensions) were compared using the ClustalW program of the Lasergene DNAstar software. Numbers indicate the nucleotide substitutions (×100). The sequences on the right depict the conserved sequence motif G-ST, at which the precursor protein is processed into the α and β polypeptides, constituting mature PS decarboxylase. The dashed line separates the two groups of PS decarboxylases containing the mitochondrial/bacterial and endomembrane forms. At-PSD1, AY189805 (At4g16700), Arabidopsis, and Le-PSD, AY093689, tomato, Rontein et al. (2003b); Sc-PSD1, L20973, yeast, Clancey et al. (1993); Cg-PSD, P27465, Chinese hamster, Kuge et al. (1991); Bs-PSD, P39822, B. subtilis, Matsumoto et al. (1998); Ec-PSD, J03916, E. coli, Li and Dowhan (1988); At-PSD2 (At5g57190) and At-PSD3, AV527283 (At4g25970), Arabidopsis, this study; Sc-PSD2, U19910, yeast, Trotter et al. (1995). The full-length coding sequences for At-PSD2 (At5g57190) and At-PSD3 (At4g25970) were deposited in GenBank with the accession numbers EF203902 and EF203901, respectively.