NETGEM: Network Embedded Temporal GEnerative Model for gene expression data

Background

Temporal analysis of gene expression data has been limited to identifying genes whose expression varies with time and/or correlation between genes that have similar temporal profiles. Often, the methods do not consider the underlying network constraints that connect the genes. It is becoming increasingly evident that interactions change substantially with time. Thus far, there is no systematic method to relate the temporal changes in gene expression to the dynamics of interactions between them. Information on interaction dynamics would open up possibilities for discovering new mechanisms of regulation by providing valuable insight into identifying time-sensitive interactions as well as permit studies on the effect of a genetic perturbation.

Results

We present NETGEM, a tractable model rooted in Markov dynamics, for analyzing the dynamics of the interactions between proteins based on the dynamics of the expression changes of the genes that encode them. The model treats the interaction strengths as random variables which are modulated by suitable priors. This approach is necessitated by the extremely small sample size of the datasets, relative to the number of interactions. The model is amenable to a linear time algorithm for efficient inference. Using temporal gene expression data, NETGEM was successful in identifying (i) temporal interactions and determining their strength, (ii) functional categories of the actively interacting partners and (iii) dynamics of interactions in perturbed networks.

Conclusions

NETGEM represents an optimal trade-off between model complexity and data requirement. It was able to deduce actively interacting genes and functional categories from temporal gene expression data. It permits inference by incorporating the information available in perturbed networks. Given that the inputs to NETGEM are only the network and the temporal variation of the nodes, this algorithm promises to have widespread applications, beyond biological systems.

The source code for NETGEM is available from https://github.com/vjethava/NETGEM

Metabolic and transcriptional response to cofactor perturbations in Escherichia coli

Metabolic cofactors such as NADH and ATP play important roles in a large number of cellular reactions, and it is of great interest to dissect the role of these cofactors in different aspects of metabolism. Toward this goal, we overexpressed NADH oxidase and the soluble F1-ATPase in Escherichia coli to lower the level of NADH and ATP, respectively. We used a global interaction network, comprising of protein interactions, transcriptional regulation, and metabolic networks, to integrate data from transcription profiles, metabolic fluxes, and the metabolite levels. We identified high-scoring networks for the two strains. The results revealed a smaller, but denser network for perturbations of ATP level, compared with that of NADH level. The action of many global transcription factors such as ArcA, Fnr, CRP, and IHF commonly involved both NADH and ATP, whereas others responded to either ATP or NADH. Overexpressing NADH oxidase invokes response in widespread aspects of metabolism involving the redox cofactors (NADH and NADPH), whereas ATPase has a more focused response to restore ATP level by enhancing proton translocation mechanisms and repressing biosynthesis. Interestingly, NADPH played a key role in restoring redox homeostasis through the concerted activity of isocitrate dehydrogenase and UdhA transhydrogenase. We present a reconciled network of regulation that illustrates the overlapping and distinct aspects of metabolism controlled by NADH and ATP. Our study contributes to the general understanding of redox and energy metabolism and should help in developing metabolic engineering strategies in E. coli.

Using regulatory information to manipulate glycerol metabolism in Saccharomyces cerevisiae

Metabolic engineering has emerged as an attractive alternative to random mutagenesis and screening to design cell factories for industrial fermentation processes. The design of metabolic networks has been realized by gene deletions or strong overexpression of heterologous genes. There is an increasing body of evidence that indicates complete inactivation of native genes and high-level activity of heterologous enzymes may be deleterious to the cell. To moderately implement their expression, genes of interest are expressed under the control of promoters with different strengths. Constructing a promoter library is labor-intensive and requires precise quantification of the promoter strength. However, when the mechanisms of pathway regulation are known, it is possible to exploit this information to effect genetic changes efficiently. We report the implementation of this concept to reducing glycerol production during aerobic growth of Saccharomyces cerevisiae. Glycerol is produced to dispose excess cytosolic reduced nicotinamide adenine dinucleotide (NADH), and the regulating step in the pathway is mediated by glycerol 3-phosphate dehydrogenase (encoded by GPD1 and GPD2 genes). We expressed NADH oxidase in S. cerevisiae under the control of the GPD2 promoter to modulate the decrease in cytosolic NADH to the right level where the heterologous enzyme does not compete with oxidative phosphorylation while at the same time, decreasing glycerol production. This metabolic design resulted in substantially decreasing glycerol production and indeed, the excess carbon was redirected to biomass, resulting in a 14% increase in the specific growth rate. We believe that such strategies are more efficient than conventional methods and will find applications in bioprocesses.

Impact of yeast systems biology on industrial biotechnology

Systems biology is yet an emerging discipline that aims to quantitatively describe and predict the functioning of a biological system. This nascent discipline relies on the recent advances in the analytical technology (such as DNA microarrays, mass spectromety, etc.) to quantify cellular characteristics (such as gene expression, protein and metabolite abundance, etc.) and computational methods to integrate information from these measurements. The model eukaryote, Saccharomyces cerevisiae, has played a pivotal role in the development of many of these analytical and computational methods and consequently is the biological system of choice for testing new hypotheses. The knowledge gained from such studies in S. cerevisiae is proving to be extremely useful in designing metabolism that is targeted to specific industrial applications. As a result, the portfolio of products that are being produced using this yeast is expanding rapidly. We review the recent developments in yeast systems biology and how they relate to industrial biotechnology.

Metabolic impact of redox cofactor perturbations in Saccharomyces cerevisiae

Redox cofactors play a pivotal role in coupling catabolism with anabolism and energy generation during metabolism. There exists a delicate balance in the intracellular level of these cofactors to ascertain an optimal metabolic output. Therefore, cofactors are emerging to be attractive targets to induce widespread changes in metabolism. We present a detailed analysis of the impact of perturbations in redox cofactors in the cytosol or mitochondria on glucose and energy metabolism in Saccharomyces cerevisiae to aid metabolic engineering decisions that involve cofactor engineering. We enhanced NADH oxidation by introducing NADH oxidase or alternative oxidase, its ATP-mediated conversion to NADPH using NADH kinase as well as the interconversion of NADH and NADPH independent of ATP by the soluble, non-proton-translocating bacterial transhydrogenase. Decreasing cytosolic NADH level lowered glycerol production, while decreasing mitochondrial NADH lowered ethanol production. However, when these reactions were coupled with NADPH production, the metabolic changes were more moderated. The direct consequence of these perturbations could be seen in the shift of the intracellular concentrations of the cofactors. The changes in product profile and intracellular metabolite levels were closely linked to the ATP requirement for biomass synthesis and the efficiency of oxidative phosphorylation, as estimated from a simple stoichiometric model. The results presented here will provide valuable insights for a quantitative understanding and prediction of cellular response to redox-based perturbations for metabolic engineering applications.

Impact of overexpressing NADH kinase on glucose and xylose metabolism in recombinant xylose-utilizing Saccharomyces cerevisiae

During growth of Saccharomyces cerevisiae on glucose, the redox cofactors NADH and NADPH are predominantly involved in catabolism and biosynthesis, respectively. A deviation from the optimal level of these cofactors often results in major changes in the substrate uptake and biomass formation. However, the metabolism of xylose by recombinant S. cerevisiae carrying xylose reductase and xylitol dehydrogenase from the fungal pathway requires both NADH and NADPH and creates cofactor imbalance during growth on xylose. As one possible solution to overcoming this imbalance, the effect of overexpressing the native NADH kinase (encoded by the POS5 gene) in xylose-consuming recombinant S. cerevisiae directed either into the cytosol or to the mitochondria was evaluated. The physiology of the NADH kinase containing strains was also evaluated during growth on glucose. Overexpressing NADH kinase in the cytosol redirected carbon flow from CO(2) to ethanol during aerobic growth on glucose and to ethanol and acetate during anaerobic growth on glucose. However, cytosolic NADH kinase has an opposite effect during anaerobic metabolism of xylose consumption by channeling carbon flow from ethanol to xylitol. In contrast, overexpressing NADH kinase in the mitochondria did not affect the physiology to a large extent. Overall, although NADH kinase did not increase the rate of xylose consumption, we believe that it can provide an important source of NADPH in yeast, which can be useful for metabolic engineering strategies where the redox fluxes are manipulated.

Increased recombinant protein production inEscherichia coli strains with overexpressed water-forming NADH oxidase and a deleted ArcA regulatory protein

Glycolytic flux is increased and acetate production is reduced in Escherichia coli by the expression of heterologous NADH oxidase (NOX) from Streptococcus pneumoniae coupled with the deletion of the arcA gene, which encodes the ArcA regulatory protein. In this study, we examined the overproduction of a model recombinant protein in strains of E. coli expressing NOX with or without an arcA mutation. The presence of NOX or the absence of ArcA reduced acetate by about 50% and increased beta-galactosidase production by 10-20%. The presence of NOX in the arcA strain eliminated acetate production entirely in batch fermentations and resulted in a 120% increase in beta-galactosidase production.

Overflow metabolism in Escherichia coli during steady-state growth: transcriptional regulation and effect of the redox ratio

Overflow metabolism in the form of aerobic acetate excretion by Escherichia coli is an important physiological characteristic of this common industrial microorganism. Although acetate formation occurs under conditions of high glucose consumption, the genetic mechanisms that trigger this phenomenon are not clearly understood. We report on the role of the NADH/NAD ratio (redox ratio) in overflow metabolism. We modulated the redox ratio in E. coli through the expression of Streptococcus pneumoniae (water-forming) NADH oxidase. Using steady-state chemostat cultures, we demonstrated a strong correlation between acetate formation and this redox ratio. We furthermore completed genome-wide transcription analyses of a control E. coli strain and an E. coli strain overexpressing NADH oxidase. The transcription results showed that in the control strain, several genes involved in the tricarboxylic acid (TCA) cycle and respiration were repressed as the glucose consumption rate increased. Moreover, the relative repression of these genes was alleviated by expression of NADH oxidase and the resulting reduced redox ratio. Analysis of a promoter binding site upstream of the genes which correlated with redox ratio revealed a degenerate sequence with strong homology with the binding site for ArcA. Deletion of arcA resulted in acetate reduction and increased the biomass yield due to the increased capacities of the TCA cycle and respiration. Acetate formation was completely eliminated by reducing the redox ratio through expression of NADH oxidase in the arcA mutant, even at a very high glucose consumption rate. The results provide a basis for studying new regulatory mechanisms prevalent at reduced NADH/NAD ratios, as well as for designing more efficient bioprocesses.

Metabolic engineering in the -omics era: elucidating and modulating regulatory networks

The importance of regulatory control in metabolic processes is widely acknowledged, and several enquiries (both local and global) are being made in understanding regulation at various levels of the metabolic hierarchy. The wealth of biological information has enabled identifying the individual components (genes, proteins, and metabolites) of a biological system, and we are now in a position to understand the interactions between these components. Since phenotype is the net result of these interactions, it is immensely important to elucidate them not only for an integrated understanding of physiology, but also for practical applications of using biological systems as cell factories. We present some of the recent "-omics" approaches that have expanded our understanding of regulation at the gene, protein, and metabolite level, followed by analysis of the impact of this progress on the advancement of metabolic engineering. Although this review is by no means exhaustive, we attempt to convey our ideology that combining global information from various levels of metabolic hierarchy is absolutely essential in understanding and subsequently predicting the relationship between changes in gene expression and the resulting phenotype. The ultimate aim of this review is to provide metabolic engineers with an overview of recent advances in complementary aspects of regulation at the gene, protein, and metabolite level and those involved in fundamental research with potential hurdles in the path to implementing their discoveries in practical applications.

Physiological response of central metabolism inEscherichia coli to deletion of pyruvate oxidase and introduction of heterologous pyruvate carboxylase

We studied the physiological response of Escherichia coli central metabolism to the expression of heterologous pyruvate carboxylase (PYC) in the presence and absence of pyruvate oxidase (POX). These studies were complemented with expression analysis of central and intermediary metabolic genes and conventional in vitro enzyme assays to evaluate glucose metabolism at steady-state growth conditions (chemostats). The absence of POX activity reduced nongrowth-related energy metabolism (maintenance coefficient) and increased the maximum specific rate of oxygen consumption. The presence of PYC activity (i.e., with POX activity) increased the biomass yield coefficient and reduced the maximum specific oxygen consumption rate compared to the wildtype. The presence of PYC in a poxB mutant resulted in a 42% lower maintenance coefficient and a 42% greater biomass yield compared to the wildtype. Providing E. coli with PYC or removing POX increased the threshold specific growth rate at which acetate accumulation began, with an 80% reduction in acetate accumulation observed at a specific growth rate of 0.4 h-1 in the poxB-pyc+ strain. Gene expression analysis suggests utilization of energetically less favorable glucose metabolism via glucokinase and the Entner-Doudoroff pathway in the absence of functional POX, while the upregulation of the phosphotransferase glucose uptake system and several amino acid biosynthetic pathways occurs in the presence of PYC. The physiological and expression changes resulting from these genetic perturbations demonstrate the importance of the pyruvate node in respiration and its impact on acetate overflow during aerobic growth.

Succinate production in dual-phase Escherichia coli fermentations depends on the time of transition from aerobic to anaerobic conditions

We examined succinic acid production in Escherichia coli AFP111 using dual-phase fermentations, which comprise an initial aerobic growth phase followed by an anaerobic production phase. AFP111 has mutations in the pfl, ldhA, and ptsG genes, and we additionally transformed this strain with the pyc gene (AFP111/pTrc99A-pyc) to provide metabolic flexibility at the pyruvate node. Aerobic fermentations with these two strains were completed to catalog physiological states during aerobic growth that might influence succinate generation in the anaerobic phase. Activities of six key enzymes were also determined for these aerobic fermentations. From these results, six transition times based on physiological states were selected for studying dual-phase fermentations. The final succinate yield and productivity depend greatly on the physiological state of the cells at the time of transition. Using the best transition time, fermentations achieved a final succinic acid concentration of 99.2 g/l with an overall yield of 110% and productivity of 1.3 g/l h.

Effects of growth mode and pyruvate carboxylase on succinic acid production by metabolically engineered strains of Escherichia coli

Escherichia coli NZN111, which lacks activities for pyruvate-formate lyase and lactate dehydrogenase, and AFP111, a derivative which contains an additional mutation in ptsG (a gene encoding an enzyme of the glucose phophotransferase system), accumulate significant levels of succinic acid (succinate) under anaerobic conditions. Plasmid pTrc99A-pyc, which expresses the Rhizobium etli pyruvate carboxylase enzyme, was introduced into both strains. We compared growth, substrate consumption, product formation, and activities of seven key enzymes (acetate kinase, fumarate reductase, glucokinase, isocitrate dehydrogenase, isocitrate lyase, phosphoenolpyruvate carboxylase, and pyruvate carboxylase) from glucose for NZN111, NZN111/pTrc99A-pyc, AFP111, and AFP111/pTrc99A-pyc under both exclusively anaerobic and dual-phase conditions (an aerobic growth phase followed by an anaerobic production phase). The highest succinate mass yield was attained with AFP111/pTrc99A-pyc under dual-phase conditions with low pyruvate carboxylase activity. Dual-phase conditions led to significant isocitrate lyase activity in both NZN111 and AFP111, while under exclusively anaerobic conditions, an absence of isocitrate lyase activity resulted in significant pyruvate accumulation. Enzyme assays indicated that under dual-phase conditions, carbon flows not only through the reductive arm of the tricarboxylic acid cycle for succinate generation but also through the glyoxylate shunt and thus provides the cells with metabolic flexibility in the formation of succinate. Significant glucokinase activity in AFP111 compared to NZN111 similarly permits increased metabolic flexibility of AFP111. The differences between the strains and the benefit of pyruvate carboxylase under both exclusively anaerobic and dual-phase conditions are discussed in light of the cellular constraint for a redox balance.