Integrative meta-analysis reveals that most yeast proteins are very stable

Abstract

Measurements for protein half-lives in yeast Saccharomyces cerevisiae reported large discrepancies, with median values between minutes to several hours. We present a unifying analysis that provides a consistent half-life estimate, based on our re-analysis of three published and one new dataset of cells grown under similar conditions. We found that degradation of many proteins can be approximated by exponential decay. Protein disappearance was primarily driven by dilution due to cell division, with cell doubling times ranging from [~]2 to 3.5 hours across the four experiments. After adjusting for doubling time, protein half-lives increased to median values between [~]7.5 to [~]40 hours. Half-lives correlated with cell doubling time even after adjustment, implying that slow growth also slows protein degradation. All estimates were validated by multiple means and were robust to different analysis methods. Overall, protein stability correlated with abundance and showed weak enrichment for degradation signals such as degrons and disordered regions. Long-lived proteins often functioned in oxidation-reduction and amino acid synthesis. Short-lived proteins often functioned in ribosome biogenesis. Despite some overall differences in behavior, all methods were able to resolve subtle difference in half-lives of ribosomal proteins, e.g. the short lifespan of RPL10. Finally, our results help the design of future experiments: time series measurements need to cover at least two to three cell doubling times for accurate estimates, exponential decay provides a reasonable proxy for protein stability, and it can be sufficiently estimated with four measurement points.