Mycobacterium bovis BCG is the only accepted vaccine for the prevention of tuberculosis (TB) in humans.BCG is a live vaccine, and induction of immunity to TB requires productive infection of the host by BCG. However, BCG is not a satisfactory vaccine, because it fails to protect against pulmonary TB in adults. In this study, we found that BCG strains cannot utilize many naturally occurring amino acids as the sole nitrogen source for growth. This defect is caused, at least partially, by the lack of functional metabolic enzymes. All BCG strains are unable to catabolize L-alanine or D-alanine due to a frameshift mutation in the L-alanine dehydrogenase gene (ald). Some BCG strains, such as BCG-Pasteur and BCG-Frappier, cannot catabolize L-serine, apparently due to inadequate expression of L-serine deaminase (sdaA). We also found that undegraded alanine and serine inhibit the growth of BCG through blockage of glutamine synthetase. These results suggest that BCG strains are limited in nitrogen metabolic capacity and predict defects that may restrict multiplication and persistence of the live vaccine within the host.Bacille Calmette-Guérin (BCG), an attenuated strain of Mycobacterium bovis isolated in 1921, is currently administered to over 100 million newborns per year as the only available vaccine against tuberculosis (TB) (13). BCG is only efficacious as a live vaccine; killed BCG does not provide protection in animal models, and killed Mycobacterium tuberculosis in human trials provided only weak and transient protection. In clinical trials of BCG vaccination, the observed efficacy has ranged from no protection to 80% fewer cases of TB (11, 13). A better understanding of this variable efficacy should help guide BCG programs and suggest avenues for the development of novel anti-TB vaccines.Several hypotheses have been generated to explain the vaccine trial data (3). The most prominent is that exposure to environmental mycobacteria partially sensitizes the host against mycobacteria and thereby provides heterologous immunity that obscures the potential benefits of BCG vaccination (14, 15). To support this, a recent study showed that the multiplication of the Danish strain of BCG was inhibited in animals previously sensitized with environmental mycobacteria. Consequently BCG vaccination elicited only a transient immune response and afforded no protection against a subsequent TB challenge (8). A second hypothesis involves differences in the vaccine strains used in clinical trials (9). After their introduction in 1921, BCG vaccines were maintained by in vitro passage in a variety of vaccine laboratories for 4 to 5 decades, resulting in phenotypic and genetic differences between BCG strains (6, 27). The capability of each of these BCG vaccines to protect against TB is unknown, because most clinical trials were performed with vaccine strains that have not been preserved. Notably, among genetic changes in BCG strains after 1921, one observes deletions in regulatory genes (which are postulated to govern the capacity to survi...