New and improved therapeutics are needed for Bacillus anthracis, the etiological agent of anthrax. To date, antimicrobial agents have not been developed against the well-validated target dihydrofolate reductase (DHFR). In order to address whether DHFR inhibitors could have potential use as clinical agents against Bacillus, 27 compounds were screened against this enzyme from Bacillus cereus, which is identical to the enzyme from B. anthracis at the active site. Several 2,4-diamino-5-deazapteridine compounds exhibit submicromolar 50% inhibitory concentrations (IC(50)s). Four of the inhibitors displaying potency in vitro were tested in vivo and showed a marked growth inhibition of B. cereus; the most potent of these has MIC(50) and minimum bactericidal concentrations at which 50% are killed of 1.6 mug/ml and 0.09 mug/ml, respectively. In order to illustrate structure-activity relationships for the classes of inhibitors tested, each of the 27 inhibitors was docked into homology models of the B. cereus and B. anthracis DHFR proteins, allowing the development of a rationale for the inhibition profiles. A combination of favorable interactions with the diaminopyrimidine and substituted phenyl rings explains the low IC(50) values of potent inhibitors; steric interactions explain higher IC(50) values. These experiments show that DHFR is a reasonable antimicrobial target for Bacillus anthracis and that there is a class of inhibitors that possess sufficient potency and antibacterial activity to suggest further development.