A Metabolic Labeling Strategy for Relative Protein Quantification in Clostridioides difficile

Clostridioides difficile (formerly Clostridium difficile) is a Gram-positive, anaerobe, spore-forming pathogen, which causes drug-induced diseases in hospitals worldwide. A detailed analysis of the proteome may provide new targets for drug development or therapeutic strategies to combat this pathogen. The application of metabolic labeling (ML) would allow for accurate quantification of significant differences in protein abundance, even in the case of very small changes. Additionally, it would be possible to perform more accurate studies of the membrane or surface proteomes, which usually require elaborated sample preparation. Such studies are therefore prone to higher standard deviations during the quantification. The implementation of ML strategies for C. difficile is complicated due to the lack in arginine and lysine auxotrophy as well as the Stickland dominated metabolism of this anaerobic pathogen. Hence, quantitative proteome analyses could only be carried out by label free or chemical labeling methods so far. In this paper, a ML approach for C. difficile is described. A cultivation procedure with (15)N-labeled media for strain 630Δerm was established achieving an incorporation rate higher than 97%. In a proof-of-principle experiment, the performance of the ML approach in C. difficile was tested. The proteome data of the cytosolic subproteome of C. difficile cells grown in complex medium as well as two minimal media in the late exponential and early stationary growth phase obtained via ML were compared with two label free relative quantification approaches (NSAF and LFQ). The numbers of identified proteins were comparable within the three approaches, whereas the number of quantified proteins were between 1,110 (ML) and 1,861 (LFQ) proteins. A hierarchical clustering showed clearly separated clusters for the different conditions and a small tree height with ML approach. Furthermore, it was shown that the quantification based on ML revealed significant altered proteins with small fold changes compared to the label free approaches. The quantification based on ML was accurate, reproducible, and even more sensitive compared to label free quantification strategies.