Many flowering plants accumulate the compatible osmolyte glycine betaine in response to osmotic stress, in certain cases together with its sulfonium analog 3-dimethylsulfoniopropionate (DMSP). Compared to glycine betaine, this DMSP accumulation appears to be an evolutionary novelty. The final step in the synthesis of glycine betaine and DMSP is oxidation of the corresponding aldehyde, betaine aldehyde or 3-dimethylsulfoniopropionaldehyde (DMSPald). Leaves of amaranth (Amaranthus hypochondriacus L.) accumulate glycine betaine but do not produce detectable amounts of DMSP. These leaves contain a betaine-aldehyde dehydrogenase (BADH) that catalyzes the final step in glycine betaine synthesis. Here we report that this enzyme efficiently catalyzes the oxidation of DMSPald. On the basis of Vmax/Km values, DMSPald is a better substrate for amaranth BADH than betaine aldehyde itself. The kinetic mechanism followed by amaranth BADH at low concentrations of DMSPald is similar to that with betaine aldehyde; as determined from initial velocity, product, dead-end, and substrate inhibition studies, it is a steady-state bi bi with ordered addition of substrates and random release of products. Like betaine aldehyde, DMSPald is inhibitory at high concentrations, at which a slower alternate route of substrate addition takes place. Our results indicate that the evolution of DMSP biosynthesis in flowering plants could have been facile in that it required no new aldehyde dehydrogenase; BADH may simply have been recruited for a novel function.