Biochemical characterization of avirulent exoC mutants of Agrobacterium tumefaciens.

The synthesis of periplasmic beta(1-2)glucan is required for crown gall tumor formation by Agrobacterium tumefaciens and for effective nodulation of alfalfa by Rhizobium meliloti. The exoC (pscA) gene is required for this synthesis by both bacteria as well as for the synthesis of capsular polysaccharide and normal lipopolysaccharide. We tested the possibility that the pleiotropic ExoC phenotype is due to a defect in the synthesis of an intermediate common to several polysaccharide biosynthetic pathways. Cytoplasmic extracts from wild-type A. tumefaciens and from exoC mutants of A. tumefaciens containing a cloned wild-type exoC gene synthesized in vitro UDP-glucose from glucose, glucose 1-phosphate, and glucose 6-phosphate. Extracts from exoC mutants synthesized UDP-glucose from glucose 1-phosphate but not from glucose or glucose 6-phosphate. Membranes from exoC mutant cells synthesized beta(1-2)glucan in vitro when exogenous UDP-glucose was added and contained the 235-kilodalton protein, which has been shown to carry out this synthesis in wild-type cells. We conclude that the inability of exoC mutants to synthesize beta(1-2)glucan is due to a deficiency in the activity of the enzyme phosphoglucomutase (EC 2.7.5.1), which in wild-type bacteria converts glucose 6-phosphate to glucose 1-phosphate, an intermediate in the synthesis of UDP-glucose. This interpretation can account for all of the deficiencies in polysaccharide synthesis which have been observed in these mutants.

Identification of the product of an Agrobacterium tumefaciens chromosomal virulence gene.

The chvB operon of Agrobacterium tumefaciens is required for bacterial attachment to plant cells and for efficient crown gall tumor formation. As defined by the virulence phenotypes of mutants with transposon insertions mapping in the region, the operon was previously mapped to a 5-kilobase (kb) stretch of chromosomal DNA. We report here that the operon is actually about 8.5 kb long and that it contains a 7-kb gene coding for a large membrane protein involved in the synthesis of cyclic beta-1,2-glucan. Mutants with transposon insertions within the 5-kb phenotypically defined operon do not synthesize this functional protein, do not synthesize beta-1,2-glucan, and do not form tumors. However, mutants with insertions that map up to 3.5 kb downstream of the phenotypically defined operon synthesize truncated proteins that are active in beta-1,2-glucan synthesis. These mutants form tumors. The truncated proteins correspond closely in size with the map positions of the insertions, suggesting that the insertions truncate the proteins by translational termination. A plasmid that contains only the phenotypically defined chvB operon also codes for a truncated protein. A fusion product between the protein and beta-galactosidase carried on a Tn3-HoHo1 insertion was observed in one mutant. Partial trypsin digestion of wild-type inner membranes generated truncated proteins that were active in beta-1,2-glucan synthesis, demonstrating that a large portion of the protein is not required for beta-1,2-glucan synthesis. The correlation between beta-1,2-glucan synthesis by the truncated proteins and tumorigenesis strongly implicates the polysaccharide product of this protein in tumor formation.