The protein expression profiles of Rhizobium leguminosarum strains in response to specific genetic perturbations in exopolysaccharide (EPS) biosynthesis genes were examined using two-dimensional gel electrophoresis. Lesions in either pssA, pssD, or pssE of R. leguminosarum bv. viciae VF39 or in pssA of R. leguminosarum bv. trifolii ANU794 not only abolished the capacity of these strains to synthesize EPS but also had a pleiotropic effect on protein synthesis levels. A minimum of 22 protein differences were observed for the two pssA mutant strains. The differences identified in the pssD and pssE mutants of strain VF39 were a distinct subset of the same protein synthesis changes that occurred in the pssA mutant. The pssD and pssE mutant strains shared identical alterations in the proteins synthesized, suggesting that they share a common function in the biosynthesis of EPS. In contrast, a pssC mutant that produces 38% of the EPS level of the parental strain showed no differences in its protein synthesis patterns, suggesting that the absence of EPS itself was contributing to the changes in protein synthesis and that there may be a complex interconnection of the EPS biosynthetic pathway with other metabolic pathways. Genetic complementation of pssA can restore wild-type protein synthesis levels, indicating that many of the observed differences in protein synthesis are also a specific response to a dysfunctional PssA. The relevance of these proteins, which are grouped as members of the pssA mutant stimulon, remains unclear, as the majority lacked a homologue in the current sequence databases and therefore possibly represent a novel functional network(s). These findings have illustrated the potential of proteomics to reveal unexpected higher-order processes of protein function and regulation that arise from mutation. In addition, it is evident that enzymatic pathways and regulatory networks are more interconnected and more sensitive to structural changes in the cell than is often appreciated. In these cases, linking the observed phenotype directly to the mutated gene can be misleading, as the phenotype could be attributable to downstream effects of the mutation.