Abstract 1422 Merits of deep and fast proteomics for bacterial proteome profiling

Bottom-up proteomics in synergy with big data science has significantly advanced microbiological research. Comprehensive proteome profiling is pivotal in advancing our knowledge of bacteria and their interactions with the environment, other organisms, and human health. Although many proteomic methods exist (sample preparation, mass-spectrometric data acquisition, and data analysis), most have only been validated on eukaryotic samples. However, bacteria possess very distinctive characteristics. For example, rigid cell walls and pathogenicity factors affect the choice of suitable cell lysis strategies. In addition to these challenges, bacteria offer certain advantages over eukaryotes: A smaller dynamic range of protein expression and a smaller genome enables faster liquid-chromatography coupled to tandem mass-spectrometry methods at constant proteomic depth. To exploit the full potential of proteomics on bacterial samples, we evaluated established protocols concerning i) the inactivation efficiency of a diverse set of bacteria, ii) suitability for high-throughput sample preparation at minimized costs, iii) the number of peptide and protein identifications, and iv) reliable quantification. We worked with six representative species to evaluate the general applicability of the workflows. The optimal sample preparation protocol combines cell lysis with 100% TFA, in-solution protein digestion, and chaotropic dilution with solid-phase extraction. Data acquisition was performed over a 30-minute linear gradient with increasing acetonitrile concentrations on an Orbitrap Exploris 480 operated in DIA mode. Data analysis was performed with DIA-NN to achieve maximal peptide and protein identifications. With this setup, we detected, on average, 40% of all open reading frames as validated for the six representative species and another 18 highly diverse bacterial species. The proposed workflow is easy to perform, robust, reproducible, and tailored to large-scale microbiological projects. We present the merits of the proposed workflow on four examples: First, comparative proteome analyses of Escherichia coli and Bacillus cereus highlight metabolic adaptions to changes in cultivation conditions. Second, the characterization of baseline proteomes of 305 bacterial species provides experimental evidence for over 500,000 genes. Additionally, this comprehensive dataset is a valuable resource for the microbiological community and will be hosted on ProteomicsDB (https://www.proteomicsdb.org/). Third, we describe non-canonical proteins in the model organism Pseudomonas aeruginosa, e.g., proteins encoded by overlapping embedded genes. Last, our workflow also supports the generation of label-free absolute quantitative data, which were used for mathematical models that explain enzyme expression kinetics during the transition from rich to minimal medium in Escherichia coli. These illustrative examples demonstrate the potential of proteomics as a versatile tool for a wide range of applications in prokaryotic research. This research was supported by EPIC-XS funded by the Horizon 2020 program of the European Union (project number 823839). T.H. received funding from NSF Grant MCB 1818384 and NIH Grant R01GM109069. A.S. was supported by Elitenetzwerk Bayern (grant number F-6-M5613.6.K-NW-2021-411/1/1).