Antibiotic resistance is a growing threat, underscoring the need to understand the underlying mechanisms. Aminoglycosides kill bacteria by disrupting translation fidelity, leading to the synthesis of aberrant proteins. Surprisingly, mutations in fusA, a gene encoding translation elongation factor G (EF-G), frequently confer resistance, even though EF-G neither participates in mRNA decoding nor blocks aminoglycoside binding. Here, we show that EF-G resistance variants selectively slow ribosome movement along mRNA when aminoglycoside is bound. This delay increases the chance that the drug dissociates before misreading occurs. Over several elongation cycles, this selective silencing of drug-bound ribosomes prevents error clusters formation, preserving proteome and membrane integrity. As a result, fusA mutations confer resistance early in treatment by preventing self-promoted aminoglycoside uptake. Translation on drug-free ribosomes remains sufficiently rapid to sustain near-normal bacterial growth. This previously unrecognized resistance mechanism—selective silencing of corrupted targets—reveals a novel antibiotic resistance strategy with potential therapeutic implications.

In addition to the analysis of misreading events by quantification of missense peptide (deposited here), the project also includes proteomics datasets, which were deposited to the ProteomeXchange Consortium via the PRIDE repository (PXD061583). For details on the terminology and quantification of errors and error clusters see also: (Translation error clusters induced by aminoglycoside antibiotics, PMID: 33758186). An overview of all deposited data, including individual acquisition parameters, can be found in the source data file of the manuscript.

Graphic created in BioRender (https://BioRender.com/40melak)