Mechanism of bactericidal action of aminoglycosides

Abstract

Since the discovery of the first aminoglycoside antibiotic, streptomycin (Str), in 1944, the mechanism of action of this group has been studied more extensively, and in more laboratories, than that of any other. The results have revealed a remarkably pleiotropic set of effects, of which four appear to be especially important: ribosomal blockade, misreading in translation, membrane damage, and irreversible uptake of the antibiotic. In contrast to our understanding of the action of the β-lactams and many other antibiotics, investigators of the aminoglycosides have struggled with a succession of hypotheses that could each account for some observations but were contradicted by others. In particular, the link between action on the ribosome and membrane damage was obscure. About 10 years ago this problem became accessible to study, with the demonstration that bacteria, like eucaryotic cells, possess membrane-bound ribosomes which are involved in protein translocation into or across the membrane. Nevertheless, the relation of this process to membrane damage by aminoglycosides was ignored until very recently, when it was proposed that incorporation of misread protein in the membrane would make it leaky to small molecules. This simple explanation suggested an obvious multistep mechanism for the bactericidal action of the aminoglycosides, in which each of the four pleiotropic effects listed above plays an equally essential role. The paper that proposed this explanation also presented supporting experimental evidence. However, its finding, that Str causes protein destined for the periplasm to accumulate in the membrane, confirms only one novel component in the multistep model: the prediction that Str induced misreading will interfere with secretion of proteins across the membrane. The other component, the hypothesis that this incorporation of misread protein in the membrane impairs its integrity, derived its support from its ability to explain a large number of already known features of aminoglycoside action. This review will now examine, in greater detail than was appropriate in the preceding experimental paper, the ability of the proposed model to accommodate all major features of aminoglycoside action.