Proteomics and constraint-based modelling reveal enzyme kinetic properties of Chlamydomonas reinhardtii on a genome scale
Biofuels produced from microalgae offer a promising solution for carbon neutral economy, and integration of turnover numbers into metabolic models can improve the design of metabolic engineering strategies towards achieving this aim. However, the coverage of enzyme turnover numbers for Chlamydomonas reinhardtii, a model eukaryotic microalga accessible to metabolic engineering, is 17-fold smaller compared to the heterotrophic model Saccharomyces cerevisiae often used as a cell factory. Here we generated protein abundance data from Chlamydomonas reinhardtii cells grown in various experiments, covering between 2337 and 3708 proteins, and employed these data with constraint-based metabolic modeling approaches to estimate in vivo maximum apparent turnover numbers for this model organism. The gathered data allowed us to estimate maximum apparent turnover numbers for 568 reactions, of which 46 correspond to transporters that are otherwise difficult to characterize. The resulting, largest-to-date catalogue of proxies for in vivo turnover numbers increased the coverage for C. reinhardtii by more than 10-fold. We showed that incorporation of these in vivo turnover numbers into a protein-constrained metabolic model of C. reinhardtii improves the accuracy of predicted enzyme usage in comparison to predictions resulting from the integration on in vitro turnover numbers. Together, the integration of proteomics and physiological data allowed us to extend our knowledge of previously uncharacterized enzymes in the C. reinhardtii genome and subsequently increase predictive performance for biotechnological applications. Significance statementCurrent metabolic modelling approaches rely on the usage of in vitro turnover numbers (kcat) that provide limited information on enzymes operating in their native environment. This knowledge gap can be closed by data-integrative approaches to estimate in vivo kcat values that can improve metabolic modelling and design of metabolic engineering strategies. In this work, we assembled a high-quality proteomics data set containing 27 samples of various culture conditions and strains of Chlamydomonas reinhardtii. We used this resource to create the largest data set of estimates for in vivo turnover numbers to date. Subsequently, we showed that metabolic models parameterized with these estimates provide better predictions of enzyme abundance than those obtained by using in vitro turnover numbers.