A Computational Model for Determining Labeling Duration in Protein Turnover Studies Using a Single Deuterated Water Labeled Sample

The time course of metabolic labeling using deuterated water, followed by liquid chromatography coupled with mass spectrometry, is employed to investigate the turnover rates of individual proteins in vivo. These labeling experiments are resource intensive. Computational methods that can determine turnover rates from a single and labeled for a short duration sample will help reduce these demands. We evaluated linear and logarithmic models to estimate protein turnover rates based on two samples (one nonlabeled and one labeled). Key factors such as the number of exchangeable hydrogens, body water enrichment in deuterium, protein turnover rate, and necessary changes in monoisotopic relative abundance established a range of labeling durations for the two-sample approach. We provide two inequalities that formally define this range of labeling duration for each peptide, which is integrated into an R Shiny App. We applied this two-sample approach to four murine tissues. By adjusting the labeling duration according to the turnover of the tissue proteome, the two-sample approach was able to analyze over 60% (1221 murine liver proteins) of the proteome previously assessed using a multisample approach.