Towards a genome-scale metabolic model of the Streptococcus pneumoniae R6 avirulent strain

The reconstruction of genome-scale metabolic models (GSMMs) is increasingly used in drug-target discovery and analysis of the secondary metabolism for the identification of potential target sites or over-production of compounds of interest [1][3]. In a general sense, genome-scale metabolic models provide insight on metabolic conversions based on genomic information and allow analysis of metabolic pathways. Streptococcus pneumoniae is a major cause of life threatening infections, such as pneumonia and meningitis, and also of less severe infections such as sinusitis and middle-ear infections. Paradoxically, it is carried asymptomatically in the human nasopharynx. It is also an important model organism within the Streptococcus genus. Although somewhat limited, biological information regarding S. pneumoniae has been increasing [4][8]. In this work, we propose a systems biology approach to further increase the S. pneumoniae knowledgebase via the reconstruction of a GSMM using merlin [9], a java framework developed in-house. The GSMMs reconstruction process is thoroughly described in literature [10], [11] and it starts with the annotation of the genome. Usually, homologous genes are found in in closely related organisms, thus a study of the phylogenetic tree of several Streptococcus species was performed to determine a reference for the annotation pipeline. merlin was used to perform the annotation of the genome taking into account the annotation of the homologous genes found with BLAST [12]and HMMER [13]. Afterwards, merlin used metabolic information from KEGG (enzymes, metabolites and reactions) to build the backbone of the network. Additionally, tool developed within the merlin framework, aimed at annotating transport proteins encoding genes (TRIAGE), generated transport reactions which were included in the draft network. Lastly, merlin also provides a tool that automatically assigns Gene-Protein-Reaction associations to enzymatic reactions in the network. The determination of the biomass equation and other constraints, as well as the formulation of in silico environmental conditions was performed by analyzing available literature. The first draft of the model encompasses 675 reactions (including 175 transport reactions), 647 metabolic species and 385 genes. The preliminary results of this model regarding gene essentiality are fairly good, as it can mimic most knockouts which present phenotypic consequences. This model will be used to assist on the study of the molecular mechanisms involved in virulence by studying the network topology and gene essentiality in different strains