The human gut microbiota plays a central role in human well-being and disease. In this study, we present an integrated, iterative approach of computational modeling, in vitro experiments, metabolomics, and genomic analysis to accelerate the identification of metabolic capabilities for poorly characterized (anaerobic) microorganisms. We demonstrate this approach for the beneficial human gut microbe Faecalibacterium prausnitzii strain A2-165. We generated an automated draft reconstruction, which we curated against the limited biochemical data. This reconstruction modeling was used to develop in silico and in vitro a chemically defined medium (CDM), which was validated experimentally. Subsequent metabolomic analysis of the spent medium for growth on CDM was performed. We refined our metabolic reconstruction according to in vitro observed metabolite consumption and secretion and propose improvements to the current genome annotation of F. prausnitzii A2-165. We then used the reconstruction to systematically characterize its metabolic properties. Novel carbon source utilization capabilities and inabilities were predicted based on metabolic modeling and validated experimentally. This study resulted in a functional metabolic map of F. prausnitzii, which is available for further applications. The presented workflow can be readily extended to other poorly characterized and uncharacterized organisms to yield novel biochemical insights about the target organism.

Innate immunity is the first line of defense against invasion of pathogens. Toll-like receptor (TLR) signaling is involved in a variety of human diseases extending far beyond immune system-related diseases, affecting a number of different tissues and cell-types. Computational models often do not account for cell-type specific differences in signaling networks. Investigation of these differences and its phenotypic implications could increase understanding of cell signaling and processes such as inflammation. The wealth of knowledge for TLR signaling has been recently summarized in a stoichiometric signaling network applicable for constraint-based modeling and analysis (COBRA). COBRA methods have been applied to investigate tissue-specific metabolism using omics data integration. Comparable approaches have not been conducted using signaling networks. In this study, we present ihsTLRv2, an updated TLR signaling network accounting for the association of 314 genes with 558 network reactions. We present a mapping procedure for transcriptomic data onto signaling networks and demonstrate the generation of a monocyte-specific TLR network. The generated monocyte network is characterized through expression of a specific set of isozymes rather than reduction of pathway contents. While further tailoring the network to a specific stimulation condition, we observed that the quantitative changes in gene expression due to LPS stimulation affected the tightly connected set of genes. Differential expression influenced about one third of the entire TLR signaling network, in particular, NF-κB activation. Thus, a cell-type and condition-specific signaling network can provide functional insight into signaling cascades. Furthermore, we demonstrate the energy dependence of TLR signaling pathways in monocytes.

The metabolic network of H. sapiens and many other organisms is described in multiple pathway databases. The level of agreement between these descriptions, however, has proven to be low. We can use these different descriptions to our advantage by identifying conflicting information and combining their knowledge into a single, more accurate, and more complete description. This task is, however, far from trivial.

Metabolic reconstructions (MRs) are common denominators in systems biology and represent biochemical, genetic, and genomic (BiGG) knowledge-bases for target organisms by capturing currently available information in a consistent, structured manner. Salmonella enterica subspecies I serovar Typhimurium is a human pathogen, causes various diseases and its increasing antibiotic resistance poses a public health problem.

Inborn errors of metabolism (IEMs) are hereditary metabolic defects, which are encountered in almost all major metabolic pathways occurring in man. Many IEMs are screened for in neonates through metabolomic analysis of dried blood spot samples. To enable the mapping of these metabolomic data onto the published human metabolic reconstruction, we added missing reactions and pathways involved in acylcarnitine (AC) and fatty acid oxidation (FAO) metabolism. Using literary data, we reconstructed an AC/FAO module consisting of 352 reactions and 139 metabolites. When this module was combined with the human metabolic reconstruction, the synthesis of 39 acylcarnitines and 22 amino acids, which are routinely measured, was captured and 235 distinct IEMs could be mapped. We collected phenotypic and clinical features for each IEM enabling comprehensive classification. We found that carbohydrate, amino acid, and lipid metabolism were most affected by the IEMs, while the brain was the most commonly affected organ. Furthermore, we analyzed the IEMs in the context of metabolic network topology to gain insight into common features between metabolically connected IEMs. While many known examples were identified, we discovered some surprising IEM pairs that shared reactions as well as clinical features but not necessarily causal genes. Moreover, we could also re-confirm that acetyl-CoA acts as a central metabolite. This network based analysis leads to further insight of hot spots in human metabolism with respect to IEMs. The presented comprehensive knowledge base of IEMs will provide a valuable tool in studying metabolic changes involved in inherited metabolic diseases.

Metabolic network reconstructions define metabolic information within a target organism and can therefore be used to address incomplete metabolic information. In the present study we used a computational approach to identify human metabolites whose metabolism is incomplete on the basis of their detection in humans but exclusion from the human metabolic network reconstruction RECON 1. Candidate solutions, composed of metabolic reactions capable of explaining the metabolism of these compounds, were then identified computationally from a global biochemical reaction database. Solutions were characterized with respect to how metabolites were incorporated into RECON 1 and their biological relevance. Through detailed case studies we show that biologically plausible non-intuitive hypotheses regarding the metabolism of these compounds can be proposed in a semi-automated manner, in an approach that is similar to de novo network reconstruction. We subsequently experimentally validated one of the proposed hypotheses and report that C9orf103, previously identified as a candidate tumour suppressor gene, encodes a functional human gluconokinase. The results of the present study demonstrate how semi-automatic gap filling can be used to refine and extend metabolic reconstructions, thereby increasing their biological scope. Furthermore, we illustrate how incomplete human metabolic knowledge can be coupled with gene annotation in order to prioritize and confirm gene functions.

Antibiotic resistance is an increasing problem in the health care system and we are in a constant race with evolving bacteria. Biofilm-associated growth is thought to play a key role in bacterial adaptability and antibiotic resistance. We employed a systems biology approach to identify candidate drug targets for biofilm-associated bacteria by imitating specific microenvironments found in microbial communities associated with biofilm formation. A previously reconstructed metabolic model of Pseudomonas aeruginosa (PA) was used to study the effect of gene deletion on bacterial growth in planktonic and biofilm-like environmental conditions. A set of 26 genes essential in both conditions was identified. Moreover, these genes have no homology with any human gene. While none of these genes were essential in only one of the conditions, we found condition-dependent genes, which could be used to slow growth specifically in biofilm-associated PA. Furthermore, we performed a double gene deletion study and obtained 17 combinations consisting of 21 different genes, which were conditionally essential. While most of the difference in double essential gene sets could be explained by different medium composition found in biofilm-like and planktonic conditions, we observed a clear effect of changes in oxygen availability on the growth performance. Eight gene pairs were found to be synthetic lethal in oxygen-limited conditions. These gene sets may serve as novel metabolic drug targets to combat particularly biofilm-associated PA. Taken together, this study demonstrates that metabolic modeling of human pathogens can be used to identify oxygen-sensitive drug targets and thus, that this systems biology approach represents a powerful tool to identify novel candidate antibiotic targets.

Host-microbe interactions play a crucial role in human health and disease. Of the various systems biology approaches, reconstruction of genome-scale metabolic networks combined with constraint-based modeling has been particularly successful at in silico predicting the phenotypic characteristics of single organisms. Here, we summarize recent studies, which have applied this approach to investigate microbe-microbe and host-microbe metabolic interactions. This approach can be also expanded to investigate the properties of an entire microbial community, as well as single organisms within the community. We illustrate that the constraint-based modeling approach is suitable to model host-microbe interactions at molecular resolution and will enable systematic investigation of metabolic links between the human host and its microbes. Such host-microbe models, combined with experimental data, will ultimately further our understanding of how microbes influence human health.

Metabolic network reconstructions formalize our knowledge of metabolism. Gaps in these networks pinpoint regions of metabolism where biological components and functions are "missing." At the same time, a major challenge in the post genomic era involves characterisation of missing biological components to complete genome annotation.

Biological systems are inherently hierarchal and multiscale in time and space. A major challenge of systems biology is to describe biological systems as a computational model, which can be used to derive novel hypothesis and drive experiments leading to new knowledge. The constraint-based reconstruction and analysis approach has been successfully applied to metabolism and to the macromolecular synthesis machinery assembly. Here, we present the first integrated stoichiometric multiscale model of metabolism and macromolecular synthesis for Escherichia coli K12 MG1655, which describes the sequence-specific synthesis and function of almost 2000 gene products at molecular detail. We added linear constraints, which couple enzyme synthesis and catalysis reactions. Comparison with experimental data showed improvement of growth phenotype prediction with the multiscale model over E. coli's metabolic model alone. Many of the genes covered by this integrated model are well conserved across enterobacters and other, less related bacteria. We addressed the question of whether the bias in synonymous codon usage could affect the growth phenotype and environmental niches that an organism can occupy. We created two classes of in silico strains, one with more biased codon usage and one with more equilibrated codon usage than the wildtype. The reduced growth phenotype in biased strains was caused by tRNA supply shortage, indicating that expansion of tRNA gene content or tRNA codon recognition allow E. coli to respond to changes in codon usage bias. Our analysis suggests that in order to maximize growth and to adapt to new environmental niches, codon usage and tRNA content must co-evolve. These results provide further evidence for the mutation-selection-drift balance theory of codon usage bias. This integrated multiscale reconstruction successfully demonstrates that the constraint-based modeling approach is well suited to whole-cell modeling endeavors.