Search

Diaminopimelic acid is an essential amino acid in the peptidoglycan of Gram-negative bacteria, mycobacteria and some Gram-positive bacteria. It can be used as a target for the development of antibacterial agents. In this article the synthesis of a series of phosphonic acid analogues of this diamino-dicarboxylic acid is described: mono- and diphosphonic acid derivatives as well as their homologues, N -alkyl analogues and some peptidyl derivatives. A phenylphosphinic derivative was also prepared. All the prepared compounds were tested for their antibacterial activity against a series of Gram-positive, Gram-negative and mycobacteria. They were also tested as possible inhibitors of the UDP-MurNAc- l -Ala-γ- d -Glu: meso -DAP synthetase. No significant biological activity was found.

Reactions of silylated phosphorus acid esters with N-triphenylmethylaldimines (N-tritylaldimines) were investigated. N-Tritylmethaneimine reacts at room temperature with a mixture of P(OMe) 3 and Me 3 SiBr forming the corresponding aminophosphonate derivatives in high yield. Other N-tritylimines are resistant toward the silylated reagents at room temperature, but undergo a similar phosphonylation reaction at elevated temperatures to form the expected aminophosphonic acids.

Retention indices for 18 monomeric stationary-phase solvents and 33 organic solutes are factor analyzed. Six factors reproduce the data matrix near experimental error. The solvents diglycerol and zonyl E-7 are shown to be highly unique, whereas none of the solutes behaves uniquely. Combination sets of solvent vectors and of solute vectors from the data matrix lead to models having errors within experimental error. Basic factors which might influence the solute—solvent interactions are identified by the least-squares method of target testing. For the solvents, a number of properties, including McReynolds' constants and molecular weight, test successfully. For the solutes, carbon number and group dipole moment are representative properties which test successfully. The best combination of solvent basic factors gives errors about twice the experimental error. Combination sets composed only of McReynolds' constants do not model the data well.