(796h) Label-Free, Strain-Resolved, Shotgun Proteomics of a Defined Bacterial Co-Culture

Systems biology approaches can provide new, detailed insights into the interactions among members of communities of microorganisms that are unavailable through conventional methods. An “-omics” approach provides metabolic information about the effects of interactions on different community members, thereby suggesting functional mechanisms to explain dynamics of the system as a whole. Recently, analysis of defined co-culture systems has been identified as a horizon for metabolic modeling and bioprocessing. Strain-resolved proteomics can track proteome changes of each species of a co-culture. To evaluate the complexity of molecular-level responses associated with growth in community, a quantitative, shotgun, label-free proteomics study was conducted on a defined bacterial co-culture consisting of two well-characterized soil species: Gram-negative Pseudomonas putida KT2440 and Gram-positive Bacillus atrophaeus ATCC 9372, the standard surrogate for the pathogen Bacillus anthracis. A total of 1151 proteins were identified across pure and co-culture samples. Differentially-expressed proteins were quantified by two methods: precursor ion intensity and spectral counts. Proteins were categorized according to Gene Ontology categories and KEGG pathways. Twenty-seven B. atrophaeus proteins and 50 P. putida proteins were significantly more abundant in the co-culture than in the respective pure culture (p<0.05; fold-change > 2), while 137 B. atrophaeus and 42 P.putida proteins were significantly less abundant in the co-culture, compared to the pure culture (p<0.05; fold-change < 2). For both species, the most common Gene Ontology biological process category for upregulated proteins was regulation of transcription. For B. atrophaeus, nearly 15% of the proteins upregulated in coculture were related to stress response. These proteins included GrpE (4. fold higher in co-culture compared to pure culture), penicillin-binding lipoprotein 3 (2 fold higher in co-culture) and glycosyltransferase (3.7 fold higher in co-culture), the latter two of which are a response to antibiotic or toxin. For P.putida, several proteins were related to toxin biosynthesis, including PPM/P mutase (2.2 fold higher in co-culture than in pure culture), 2-dehydro-3-deoxyphosphooctonate aldolase 2 (2. fold higher in co-culture), phenazine biosynthesis protein (6.5 fold higher than in co-culture), and GTP cyclohydrolase-2 (2.3 fold higher in co-culture). Futhermore, for each species several differentially-expressed proteins related to iron uptake. For B.atrophaeus, 2,3-dihydroxybenzoate-AMP ligase, an enzyme in the siderophore biosynthesis pathway, was 7-fold higher in the co-culture than in pure culture. Bacterioferritin was down-regulated in P.putida. Follow-up growth studies involving growth in culture supernatant and at different levels of iron availability were conducted to verify interactions suggested by the proteomics data. This study contributes to a growing literature that seeks to discover and characterize patterns of microbial interactions through powerful “-omics” analysis of co-culture systems of model species.