A Pan-Genomic Guided Metabolic Reconstruction and Omics Integration to Understand Propionibacterium Metabolism

High GC content microorganisms are our richest source of antibiotics and encompass many industrial microorganisms and pathogens, including Propionibacterium. This signature high GC content of Propionibacterium has hindered both the acquisition of quality sequence data and development of genetic modification tools for the genus. As a result, we do not yet really understand the metabolism of Propionibacterium despite extensive research efforts and their relevance to industry and human health. To address this issue, a functional annotation in the form of a genome-scale model can guide our understanding of this important genus of microorganisms. Our first attempt to reconstruct a network quickly reflected the limitations of the high GC content sequence and subsequent annotation; central metabolic pathways to primary fermentation products were incomplete or absent including the characteristic Wood-Werkman cycle responsible for propionate production. Because of improvements to sequencing technology, we have seen an increasing availability of various genome sequences for Propionibacterium species. To capitalise on the limited available experimental and sequence data, we developed a pan-genomic guided metabolic reconstruction strategy to integrate this data to produce and curate the first high quality Propionibacterium genome scale model and a pan-Propionibacterium model.

With the establishment of the comparative bioinformatics and in silico platforms we could more deeply explore Propionibacterium metabolism. The combined analysis demonstrated an unexpected diversity in their central carbon metabolism and a hidden layer of complexity. Additionally, the new insights into Propionibacterium metabolism allowed us to generate several hypotheses which were tested in silico, including the existence of a novel, putative ferredoxin-linked energy conservation strategy. We further explored metabolism by coupling transcriptomics data to the genome-scale model to identify phenotypical behaviours that are inconsistent with in silico predicted optimal growth, such as those potentially associated with acid tolerance. Finally, we extended this analytical framework to investigate orthologous sequences and metabolic functionalities associated with high and low propionic acid producing Propionibacterium species [1].

Overall, our approach highlights the synergistic complementation of the pan-genomic and pan-GEM analyses which allowed us to circumvent challenges associated with and produce the first quality reconstruction of Propionibacterium. Additionally our approach allowed us to gain insight into metabolic features driving phenotypic differences between strains. We expect reconstructions of other non-model organisms and high GC content bacteria will benefit by employing a similar strategy.

 1. Stowers CC, Cox BM, Rodriguez BA: Development of an industrializable fermentation process for propionic acid production. J Ind Microbiol Biotechnol 2014, 41:837–52.