Summary Phosphorylation of the translation initiation factor eIF2α is a rapid and vital response to many forms of stress, including protein-misfolding stress in the endoplasmic reticulum (ER stress). It is believed to cause a general reduction in protein synthesis while enabling translation of few transcripts. Such a reduction of protein synthesis comes with the threat of depleting essential proteins, a risk thought to be mitigated by its transient nature. Here, we find that translation attenuation is not uniform, with cytosolic and mitochondrial ribosomal subunits being prominently downregulated. Translation attenuation of these targets persists after translation recovery. Surprisingly, this occurs without a measurable decrease in ribosomal proteins. Explaining this conundrum, translation attenuation preferentially targets long-lived proteins, a finding not only demonstrated by ribosomal proteins but also observed at a global level. This shows that protein stability buffers the cost of translational attenuation, establishing an evolutionary principle of cellular robustness.

The sulfur-containing amino acids methionine and cysteine play an important role in food industry. These amino acids are used to confer a sulfur smell or meat-related aroma to food products. Besides their use as food additives, methionine and cysteine participate in flavor formation in dairy fermentations. For instance, the characteristic aroma of Cheddar cheeses is derived from methionine. Therefore, bacterial strains with the ability to overproduce and secrete these amino acids are relevant for the food industry. In addition, the quantification of these compounds in food matrices is a laborious task that involves sample preparation and specific analytical methods such as high-performance liquid chromatography. The ability of bacteria to naturally sense metabolites has successfully been exploited to develop biosensors. The presence of a specific metabolite is sensed by the biosensors, and it is subsequently translated into the expression of one or more reporter genes. In this study we aim to develop biosensors to detect methionine and cysteine, which are produced and secreted by wild-type Lactococcus lactis strains. We employed two strategies to create L. lactis biosensors, the first one is based on the methionine auxotrophy of this bacterium and the second strategy is based on a cysteine-responsive promoter. The characterization of the biosensors showed their specific response to the presence of these amino acids. Subsequently, we applied the methionine biosensor to quantify the presence of methionine in bacterial supernatants of wild-type L. lactis that naturally secretes methionine to benchmark the performance of our biosensors. The methionine biosensor responded linearly to the amounts of methionine present in the bacterial supernatants, i.e., the increases in the biosensor cell densities were proportional to the amounts of methionine present in the supernatants. The biosensors developed in this study tackle the limitations of amino acid quantification and the selection of strains with secretion of amino acids. These biosensors may eventually be used for screening of engineered strains to increase methionine and cysteine production, and may facilitate the detection of these amino acids in complex food matrices.