Certain plants produce glycine betaine (GlyBet) in the chloroplast by a two-step oxidation of choline. Introducing GlyBet accumulation into plants that lack it is a well-established target for metabolic engineering because GlyBet can lessen damage from osmotic stress. The first step in GlyBet synthesis is catalyzed by choline monooxygenase (CMO), a stromal enzyme with a Rieske-type [2Fe-2S] center. The absence of CMO is the primary constraint on GlyBet production in GlyBet-deficient plants such as tobacco, but the endogenous choline supply is also potentially problematic. To investigate this, we constructed transgenic tobacco plants that constitutively express a spinach CMO cDNA. The CMO protein was correctly compartmented in chloroplasts and was enzymatically active, showing that its [2Fe-2S] cluster had been inserted. Salinization increased CMO protein levels, apparently via a post-transcriptional mechanism, to as high as 10% of that in salinized spinach. However, the GlyBet contents of CMO+ plants were very low (0.02-0.05 μmol g−1 fresh weight) in both unstressed and salinized conditions. Experiments with [14C]GlyBet demonstrated that this was not due to GlyBet catabolism. When CMO+ plants were supplied in culture with 5 mM choline or phosphocholine, their choline and GlyBet levels increased by at least 30-fold. The choline precursors mono- and dimethylethanolamine also enhanced choline and GlyBet levels but ethanolamine did not, pointing to a major constraint on flux to choline at the first methylation step in its synthesis. The extractable activity of the enzyme mediating this step in tobacco was only 3% that of spinach. We conclude that in GlyBet-deficient plants engineered with choline-oxidizing genes, the size of the free choline pool and the metabolic flux to choline need to be increased to attain GlyBet levels as high as those in natural accumulators.