Synthesis of beta-ketophosphonate analogs of glutamyl and glutaminyl adenylate, and selective inhibition of the corresponding bacterial aminoacyl-tRNA synthetases.

The aminoacyl-beta-ketophosphonate-adenosines (aa-KPA) are stable analogs of the aminoacyl adenylates, which are high-energy intermediates in the formation of aminoacyl-tRNA catalyzed by aminoacyl-tRNA synthetases (aaRS). We have synthesized glutamyl-beta-ketophosphonate-adenosine (Glu-KPA) and glutaminyl-beta-ketophosphonate-adenosine (Gln-KPA), and have tested them as inhibitors of their cognate aaRS, and of a non-cognate aaRS. Glu-KPA is a competitive inhibitor of Escherichia coli glutamyl-tRNA synthetase (GluRS) with a K(i) of 18microM with respect to its substrate glutamate, and binds at one site on this monomeric enzyme; the non-cognate Gln-KPA also binds this GluRS at one site, but is a much weaker (K(i)=2.9mM) competitive inhibitor. By contrast, Gln-KPA inhibits E. coli glutaminyl-tRNA synthetase (GlnRS) by binding competitively but weakly at two distinct sites on this enzyme (average K(i) of 0.65mM); the non-cognate Glu-KPA shows one-site weak (K(i)=2.8mM) competitive inhibition of GlnRS. These kinetic results indicate that the glutamine and the AMP modules of Gln-KPA, connected by the beta-ketophosphonate linker, cannot bind GlnRS simultaneously, and that one Gln-KPA molecule binds the AMP-binding site of GlnRS through its AMP module, whereas another Gln-KPA molecule binds the glutamine-binding site through its glutamine module. This model suggests that similar structural constraints could affect the binding of Glu-KPA to the active site of mammalian cytoplasmic GluRSs, which are evolutionarily much closer to bacterial GlnRS than to bacterial GluRS. This possibility was confirmed by the fact that Glu-KPA inhibits bovine liver GluRS 145-fold less efficiently than E. coli GluRS by competitive weak binding at two distinct sites (average K(i)=2.6mM). Moreover, these kinetic differences reveal that the active sites of bacterial GluRSs and mammalian cytoplasmic GluRSs have substantial structural differences that could be further exploited for the design of better inhibitors specific for bacterial GluRSs, promising targets for antimicrobial therapy.