How bacterial stress responses reshape the relationship between mRNA and protein levels

Bacterial pathogens encounter a wide range of environmental stresses during infection, including nutrient limitation, changes in temperature and pH, oxidative stress, and reduced oxygen availability. To survive these challenging conditions, bacteria activate complex regulatory mechanisms that alter gene expression and protein production. Although messenger RNA (mRNA) levels are commonly used to study bacterial responses, they do not always accurately reflect the abundance of the corresponding proteins. Understanding this relationship is therefore essential for identifying the molecular mechanisms that enable bacterial pathogens to adapt and survive in hostile host environments.

Süer et al. (2026) investigated how environmental stress affects the relationship between mRNA and protein levels in three clinically relevant bacterial pathogens: Salmonella enterica Typhimurium, Yersinia pseudotuberculosis, and Staphylococcus aureus. The researchers integrated transcriptomic data from the previously established PATHOgenex RNA Atlas with newly generated mass-spectrometry-based proteomic data. They examined bacterial responses under 10 infection-relevant stress conditions, including acidic, osmotic, oxidative, and nitrosative stress, hypoxia, nutrient limitation, and temperature changes. By comparing gene expression at both the RNA and protein levels, the study provides a comprehensive overview of how bacterial stress responses are regulated.

The analysis revealed that mRNA and protein levels were positively correlated across all three pathogens, but the strength of this relationship varied considerably under different stress conditions. Correlation coefficients ranged from 0.48 to 0.68 in S. enterica Typhimurium, 0.54 to 0.72 in S. aureus, and 0.59 to 0.76 in Y. pseudotuberculosis. Conditions that triggered strong stress responses, particularly hypoxia, nutrient limitation, and stationary phase, generally were associated with weaker relationships. Interestingly, essential genes maintained higher and more stable mRNA and protein expression levels than nonessential genes, suggesting that proteins required for fundamental cellular functions are more consistently regulated during environmental stress.

A particularly important finding was the identification of large numbers of genes whose mRNA and protein levels responded differently to stress. Across the tested conditions, approximately 32.5% of detected genes in S. enterica Typhimurium, 32.1% in S. aureus, and 26.5% in Y. pseudotuberculosis showed discordant regulation on average. In many cases, mRNA abundance changed significantly while the corresponding protein levels remained relatively stable. These differences were particularly evident among genes involved in translation, RNA processing, protein modification, and other essential cellular processes. The findings suggest that bacteria use regulatory mechanisms beyond transcription to control protein abundance and maintain cellular functions during stress.

Osmotic stress emerged as a particularly interesting condition, producing pronounced differences between mRNA and protein levels among stress-specific genes. Using the MOBILE computational framework, the researchers identified associations between RNA and protein expression linked to translation-related processes. Follow-up experiments demonstrated that osmotic stress reduced active protein synthesis in Y. pseudotuberculosis and S. enterica Typhimurium. This provides experimental evidence that impaired translation contributes to the observed mismatch between mRNA and protein abundance, even when bacterial cells continue to express stress-response genes.

By integrating transcriptomics and proteomics across multiple bacterial species and environmental conditions, Süer et al. (2026) demonstrate that bacterial stress adaptation cannot be fully understood from gene expression data alone. Their findings highlight the importance of translational and other posttranscriptional regulatory processes in shaping bacterial responses to infection-relevant environments. The resulting multispecies dataset provides a valuable resource for investigating bacterial physiology, identifying potential antimicrobial targets, and improving our understanding of how pathogens survive under challenging conditions.

Data

• The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifier PXD055243.
• The transcriptome data published for S. enterica, Y. pseudotuberculosis, and S. aureus were obtained from PATHOgenex RNA Atlas with Gene Expression Omnibus (GEO) Accession No. GSE152295.
• The complete dataset is available on Zenodo, and can be used with the associated Rnotebook.
• A repository containing an Rnotebook with all the code to reproduce the figures and analyses presented in this manuscript is available on GitHub. All packages and versions are listed in the Renv file available in the GitHub repository.

Article

DOI: https://doi.org/10.1073/pnas.2613102123

Süer, S. G., Arnoux, J., Lim, Y. Y., Dhurve, G., Şen, R., Erdem, C., Mateus, A., & Avican, K. (2026). Bacterial stress responses lower mRNA–protein level correlations. Proceedings of the National Academy of Sciences, 123(38), e2613102123.

Contact:

For further information about the study, please contact the authors at sena.gizem@umu.se (or avicanlab@gmail.com) or visit the Avican Lab webpage.