Sergio Bordel

Reconstruction of Genome-Scale Active Metabolic Networks for 69 Human Cell Types and 16 Cancer Types Using INIT

Development of high throughput analytical methods has given physicians the potential access to extensive and patient-specific data sets, such as gene sequences, gene expression profiles or metabolite footprints. This opens for a new approach in health care, which is both personalized and based on system-level analysis. Genome-scale metabolic networks provide a mechanistic description of the relationships between different genes, which is valuable for the analysis and interpretation of large experimental data-sets. Here we describe the generation of genome-scale active metabolic networks for 69 different cell types and 16 cancer types using the INIT (Integrative Network Inference for Tissues) algorithm. The INIT algorithm uses cell type specific information about protein abundances contained in the Human Proteome Atlas as the main source of evidence. The generated models constitute the first step towards establishing a Human Metabolic Atlas, which will be a comprehensive description (accessible online) of the metabolism of different human cell types, and will allow for tissue-level and organism-level simulations in order to achieve a better understanding of complex diseases. A comparative analysis between the active metabolic networks of cancer types and healthy cell types allowed for identification of cancer-specific metabolic features that constitute generic potential drug targets for cancer treatment.

Understanding the interactions between bacteria in the human gut through metabolic modeling

The human gut microbiome plays an influential role in maintaining human health, and it is a potential target for prevention and treatment of disease. Genome-scale metabolic models (GEMs) can provide an increased understanding of the mechanisms behind the effects of diet, the genotype-phenotype relationship and microbial robustness. Here we reconstructed GEMs for three key species, (Bacteroides thetaiotamicron, Eubacterium rectale and Methanobrevibacter smithii) as relevant representatives of three main phyla in the human gut (Bacteroidetes, Firmicutes and Euryarchaeota). We simulated the interactions between these three bacteria in different combinations of gut ecosystems and compared the predictions with the experimental results obtained from colonization of germ free mice. Furthermore, we used our GEMs for analyzing the contribution of each species to the overall metabolism of the gut microbiota based on transcriptome data and demonstrated that these models can be used as a scaffold for understanding bacterial interactions in the gut.

Sampling the Solution Space in Genome-Scale Metabolic Networks Reveals Transcriptional Regulation in Key Enzymes

Genome-scale metabolic models are available for an increasing number of organisms and can be used to define the region of feasible metabolic flux distributions. In this work we use as constraints a small set of experimental metabolic fluxes, which reduces the region of feasible metabolic states. Once the region of feasible flux distributions has been defined, a set of possible flux distributions is obtained by random sampling and the averages and standard deviations for each of the metabolic fluxes in the genome-scale model are calculated. These values allow estimation of the significance of change for each reaction rate between different conditions and comparison of it with the significance of change in gene transcription for the corresponding enzymes. The comparison of flux change and gene expression allows identification of enzymes showing a significant correlation between flux change and expression change (transcriptional regulation) as well as reactions whose flux change is likely to be driven only by changes in the metabolite concentrations (metabolic regulation). The changes due to growth on four different carbon sources and as a consequence of five gene deletions were analyzed for Saccharomyces cerevisiae. The enzymes with transcriptional regulation showed enrichment in certain transcription factors. This has not been previously reported. The information provided by the presented method could guide the discovery of new metabolic engineering strategies or the identification of drug targets for treatment of metabolic diseases.

Mapping condition-dependent regulation of metabolism in yeast through genome-scale modeling.

BackgroundThe genome-scale metabolic model of Saccharomyces cerevisiae, first presented in 2003, was the first genome-scale network reconstruction for a eukaryotic organism. Since then continuous efforts have been made in order to improve and expand the yeast metabolic network.ResultsHere we present iTO977, a comprehensive genome-scale metabolic model that contains more reactions, metabolites and genes than previous models. The model was constructed based on two earlier reconstructions, namely iIN800 and the consensus network, and then improved and expanded using gap-filling methods and by introducing new reactions and pathways based on studies of the literature and databases. The model was shown to perform well both for growth simulations in different media and gene essentiality analysis for single and double knock-outs. Further, the model was used as a scaffold for integrating transcriptomics, and flux data from four different conditions in order to identify transcriptionally controlled reactions, i.e. reactions that change both in flux and transcription between the compared conditions.ConclusionWe present a new yeast model that represents a comprehensive up-to-date collection of knowledge on yeast metabolism. The model was used for simulating the yeast metabolism under four different growth conditions and experimental data from these four conditions was integrated to the model. The model together with experimental data is a useful tool to identify condition-dependent changes of metabolism between different environmental conditions.

Use of genome-scale metabolic models for understanding microbial physiology.

The exploitation of microorganisms in industrial, medical, food and environmental biotechnology requires a comprehensive understanding of their physiology. The availability of genome sequences and accumulation of high‐throughput data allows gaining understanding of microbial physiology at the systems level, and genome‐scale metabolic models represent a valuable framework for integrative analysis of metabolism of microorganisms. Genome‐scale metabolic models are reconstructed based on a combination of genome sequence information and detailed biochemical information, and these reconstructed models can be used for analyzing and simulating the operation of metabolism in response to different stimuli. Here we discuss the requirement for having detailed physiological insight in order to exploit microorganisms for production of fuels, chemicals and pharmaceuticals. We further describe the reconstruction process of genome‐scale metabolic models and different algorithms that can be used to apply these models to gain improved insight into microbial physiology.

A systematic selection of the non-aqueous phase in a bacterial two liquid phase bioreactor treating α-pinene

A systematic evaluation of the selection criteria of non-aqueous phases in two liquid phase bioreactors (TLPBs), also named two-phase partitioning bioreactors (TPPBs), was carried out using the biodegradation of α-pinene by Pseudomonas fluorescens NCIMB 11671 as a model process. A preliminary solvent screening was thus carried out among the most common non-aqueous phases reported in literature for volatile organic contaminants biodegradation in TLPBs: silicon oil, paraffin oil, hexadecane, diethyl sebacate, dibutyl-phtalate, FC 40, 1,1,1,3,5,5,5-heptamethyltrisiloxane (HMS), and 2,2,4,4,6,8,8-heptamethylnonane (HMN). FC 40, silicone oil, HMS, and HMN were first selected based on its biocompatibility, resistance to microbial attack, and α-pinene mass transport characteristics. FC 40, HMS, HMN, and silicone oil at 10% (v/v) enhanced α-pinene mass transport from the gas to the liquid phase by a factor of 3.8, 14.8, 11.4, and 8.6, respectively, compared to a single-phase aqueous system. FC 40 and HMN were finally compared for their ability to enhance α-pinene biodegradation in a mechanically agitated bioreactor. The use of FC 40 or HMN (both at 10% v/v) sustained non-steady state removal efficiencies (RE) and elimination capacities (EC) approximately 7 and 12 times higher than those achieved in the system without an organic phase, respectively. In addition, preliminary results showed that P fluorescens could uptake and mineralize α-pinene directly from the non aqueous phase.

Genome-Scale Modeling of the Protein Secretory Machinery in Yeast

The protein secretory machinery in Eukarya is involved in post-translational modification (PTMs) and sorting of the secretory and many transmembrane proteins. While the secretory machinery has been well-studied using classic reductionist approaches, a holistic view of its complex nature is lacking. Here, we present the first genome-scale model for the yeast secretory machinery which captures the knowledge generated through more than 50 years of research. The model is based on the concept of a Protein Specific Information Matrix (PSIM: characterized by seven PTMs features). An algorithm was developed which mimics secretory machinery and assigns each secretory protein to a particular secretory class that determines the set of PTMs and transport steps specific to each protein. Protein abundances were integrated with the model in order to gain system level estimation of the metabolic demands associated with the processing of each specific protein as well as a quantitative estimation of the activity of each component of the secretory machinery.

Non-small cell lung cancer is characterized by dramatic changes in phospholipid profiles

Non‐small cell lung cancer (NSCLC) is the leading cause of cancer death globally. To develop better diagnostics and more effective treatments, research in the past decades has focused on identification of molecular changes in the genome, transcriptome, proteome, and more recently also the metabolome. Phospholipids, which nevertheless play a central role in cell functioning, remain poorly explored. Here, using a mass spectrometry (MS)‐based phospholipidomics approach, we profiled 179 phospholipid species in malignant and matched non‐malignant lung tissue of 162 NSCLC patients (73 in a discovery cohort and 89 in a validation cohort). We identified 91 phospholipid species that were differentially expressed in cancer versus non‐malignant tissues. Most prominent changes included a decrease in sphingomyelins (SMs) and an increase in specific phosphatidylinositols (PIs). Also a decrease in multiple phosphatidylserines (PSs) was observed, along with an increase in several phosphatidylethanolamine (PE) and phosphatidylcholine (PC) species, particularly those with 40 or 42 carbon atoms in both fatty acyl chains together. 2D‐imaging MS of the most differentially expressed phospholipids confirmed their differential abundance in cancer cells. We identified lipid markers that can discriminate tumor versus normal tissue and different NSCLC subtypes with an AUC (area under the ROC curve) of 0.999 and 0.885, respectively. In conclusion, using both shotgun and 2D‐imaging lipidomics analysis, we uncovered a hitherto unrecognized alteration in phospholipid profiles in NSCLC. These changes may have important biological implications and may have significant potential for biomarker development.

Codon usage variability determines the correlation between proteome and transcriptome fold changes

BackgroundThe availability of high throughput experimental methods has made possible to observe the relationships between proteome and transcirptome. The protein abundances show a positive but weak correlation with the concentrations of their cognate mRNAs. This weak correlation implies that there are other crucial effects involved in the regulation of protein translation, different from the sole availability of mRNA. It is well known that ribosome and tRNA concentrations are sources of variation in protein levels. Thus, by using integrated analysis of omics data, genomic information, transcriptome and proteome, we aim to unravel important variables affecting translation.ResultsWe identified how much of the variability in the correlation between protein and mRNA concentrations can be attributed to the gene codon frequencies. We propose the hypothesis that the influence of codon frequency is due to the competition of cognate and near-cognate tRNA binding; which in turn is a function of the tRNA concentrations. Transcriptome and proteome data were combined in two analytical steps; first, we used Self-Organizing Maps (SOM) to identify similarities among genes, based on their codon frequencies, grouping them into different clusters; and second, we calculated the variance in the protein mRNA correlation in the sampled genes from each cluster. This procedure is justified within a mathematical framework.ConclusionsWith the proposed method we observed that in all the six studied cases most of the variability in the relation protein-transcript could be explained by the variation in codon composition.

Mapping global effects of the anti-sigma factor MucA in Pseudomonas fluorescens SBW25 through genome-scale metabolic modeling

BackgroundAlginate is an industrially important polysaccharide, currently produced commercially by harvesting of marine brown sea-weeds. The polymer is also synthesized as an exo-polysaccharide by bacteria belonging to the genera Pseudomonas and Azotobacter, and these organisms may represent an alternative alginate source in the future. The current work describes an attempt to rationally develop a biological system tuned for very high levels of alginate production, based on a fundamental understanding of the system through metabolic modeling supported by transcriptomics studies and carefully controlled fermentations.ResultsAlginate biosynthesis in Pseudomonas fluorescens was studied in a genomics perspective, using an alginate over-producing strain carrying a mutation in the anti-sigma factor gene mucA. Cells were cultivated in chemostats under nitrogen limitation on fructose or glycerol as carbon sources, and cell mass, growth rate, sugar uptake, alginate and CO2 production were monitored. In addition a genome scale metabolic model was constructed and samples were collected for transcriptome analyses. The analyses show that polymer production operates in a close to optimal way with respect to stoichiometric utilization of the carbon source and that the cells increase the uptake of carbon source to compensate for the additional needs following from alginate synthesis. The transcriptome studies show that in the presence of the mucA mutation, the alg operon is upregulated together with genes involved in energy generation, genes on both sides of the succinate node of the TCA cycle and genes encoding ribosomal and other translation-related proteins. Strains expressing a functional MucA protein (no alginate production) synthesize cellular biomass in an inefficient way, apparently due to a cycle that involves oxidation of NADPH without ATP production. The results of this study indicate that the most efficient way of using a mucA mutant as a cell factory for alginate production would be to use non-growing conditions and nitrogen deprivation.ConclusionsThe insights gained in this study should be very useful for a future efficient production of microbial alginates.