Logistic PCA explains differences between genome-scale metabolic models in terms of metabolic pathways.

Genome-scale metabolic models (GSMMs) offer a holistic view of biochemical reaction networks, enabling in-depth analyses of metabolism across species and tissues in multiple conditions. However, comparing GSMMs Against each other poses challenges as current dimensionality reduction algorithms or clustering methods lack mechanistic interpretability, and often rely on subjective assumptions. Here, we propose a new approach utilizing logisitic principal component analysis (LPCA) that efficiently clusters GSMMs while singling out mechanistic differences in terms of reactions and pathways that drive the categorization. We applied LPCA to multiple diverse datasets, including GSMMs of 222 Escherichia-strains, 343 budding yeasts (Saccharomycotina), 80 human tissues, and 2943 Firmicutes strains. Our findings demonstrate LPCA's effectiveness in preserving microbial phylogenetic relationships and discerning human tissue-specific metabolic profiles, exhibiting comparable performance to traditional methods like t-distributed stochastic neighborhood embedding (t-SNE) and Jaccard coefficients. Moreover, the subsystems and associated reactions identified by LPCA align with existing knowledge, underscoring its reliability in dissecting GSMMs and uncovering the underlying drivers of separation.

Elementary vectors and autocatalytic sets for resource allocation in next-generation models of cellular growth.

Traditional (genome-scale) metabolic models of cellular growth involve an approximate biomass "reaction", which specifies biomass composition in terms of precursor metabolites (such as amino acids and nucleotides). On the one hand, biomass composition is often not known exactly and may vary drastically between conditions and strains. On the other hand, the predictions of computational models crucially depend on biomass. Also elementary flux modes (EFMs), which generate the flux cone, depend on the biomass reaction. To better understand cellular phenotypes across growth conditions, we introduce and analyze new classes of elementary vectors for comprehensive (next-generation) metabolic models, involving explicit synthesis reactions for all macromolecules. Elementary growth modes (EGMs) are given by stoichiometry and generate the growth cone. Unlike EFMs, they are not support-minimal, in general, but cannot be decomposed "without cancellations". In models with additional (capacity) constraints, elementary growth vectors (EGVs) generate a growth polyhedron and depend also on growth rate. However, EGMs/EGVs do not depend on the biomass composition. In fact, they cover all possible biomass compositions and can be seen as unbiased versions of elementary flux modes/vectors (EFMs/EFVs) used in traditional models. To relate the new concepts to other branches of theory, we consider autocatalytic sets of reactions. Further, we illustrate our results in a small model of a self-fabricating cell, involving glucose and ammonium uptake, amino acid and lipid synthesis, and the expression of all enzymes and the ribosome itself. In particular, we study the variation of biomass composition as a function of growth rate. In agreement with experimental data, low nitrogen uptake correlates with high carbon (lipid) storage.

Inclusion of maintenance energy improves the intracellular flux predictions of CHO.

Chinese hamster ovary (CHO) cells are the leading platform for the production of biopharmaceuticals with human-like glycosylation. The standard practice for cell line generation relies on trial and error approaches such as adaptive evolution and high-throughput screening, which typically take several months. Metabolic modeling could aid in designing better producer cell lines and thus shorten development times. The genome-scale metabolic model (GSMM) of CHO can accurately predict growth rates. However, in order to predict rational engineering strategies it also needs to accurately predict intracellular fluxes. In this work we evaluated the agreement between the fluxes predicted by parsimonious flux balance analysis (pFBA) using the CHO GSMM and a wide range of 13C metabolic flux data from literature. While glycolytic fluxes were predicted relatively well, the fluxes of tricarboxylic acid (TCA) cycle were vastly underestimated due to too low energy demand. Inclusion of computationally estimated maintenance energy significantly improved the overall accuracy of intracellular flux predictions. Maintenance energy was therefore determined experimentally by running continuous cultures at different growth rates and evaluating their respective energy consumption. The experimentally and computationally determined maintenance energy were in good agreement. Additionally, we compared alternative objective functions (minimization of uptake rates of seven nonessential metabolites) to the biomass objective. While the predictions of the uptake rates were quite inaccurate for most objectives, the predictions of the intracellular fluxes were comparable to the biomass objective function.

What CHO is made of: Variations in the biomass composition of Chinese hamster ovary cell lines.

BACKGROUND: Cell line-specific, genome-scale metabolic models enable rigorous and systematic in silico investigation of cellular metabolism. Such models have recently become available for Chinese hamster ovary (CHO) cells. However, a key ingredient, namely an experimentally validated biomass function that summarizes the cellular composition, was so far missing. Here, we close this gap by providing extensive experimental data on the biomass composition of 13 parental and producer CHO cell lines under various conditions. RESULTS: We report total protein, lipid, DNA, RNA and carbohydrate content, cell dry mass, and detailed protein and lipid composition. Furthermore, we present meticulous data on exchange rates between cells and environment and provide detailed experimental protocols on how to determine all of the above. The biomass composition is converted into cell line- and condition-specific biomass functions for use in cell line-specific, genome-scale metabolic models of CHO. Finally, flux balance analysis (FBA) is used to demonstrate consistency between in silico predictions and experimental analysis. CONCLUSIONS: Our study reveals a strong variability of the total protein content and cell dry mass across cell lines. However, the relative amino acid composition is independent of the cell line and condition and thus needs not be explicitly measured for each new cell line. In contrast, the lipid composition is strongly influenced by the growth media and thus will have to be determined in each case. These cell line-specific variations in biomass composition have a small impact on growth rate predictions with FBA, as inaccuracies in the predictions are rather dominated by inaccuracies in the exchange rate spectra. Cell-specific biomass variations only become important if the experimental errors in the exchange rate spectra drop below twenty percent.

Optimization of the fermentation and downstream processes for human enterokinase production in Pichia pastoris.

Enterokinase is one of the most frequently used enzymes for the removal of affinity tags from target recombinant proteins. In this study, several fermentation strategies were assayed for the production of human enterokinase in Pichia pastoris under constitutive GAP promoter. Two of them with controlled specific growth rate during whole cultivation showed a very low enterokinase activity, under 1 U/ml, of the fermentation medium. On the contrary, the combined fermentation with a maximum specific growth rate at the initial phase of the fermentation and stationary-like phase during the rest of the fermentation showed a significant accumulation of the enterokinase in the medium, which counted up to 1400 U/ml. Lower cultivation temperature had a negative impact on the enzyme accumulation during this fermentation strategy. Downstream processes were focused on buffer environment optimization directly after cultivation, as at this time, the most amount of the activity is eliminated by endogenous proteases. Slightly positive effect on enzyme activity in the medium had an addition of liquid storage solution of EDTA and KOH to adjust pH to 8 and molarity of the EDTA to 50 mM. During the purification process, a significant amount of the enzyme was detected to be lost, which counted up to 90%. The purified enzyme, enterokinase, kept quality standard of the published enzymes.

Costs of ribosomal RNA stabilization affect ribosome composition at maximum growth rate.

Ribosomes are key to cellular self-fabrication and limit growth rate. While most enzymes are proteins, ribosomes consist of 1/3 protein and 2/3 ribonucleic acid (RNA) (in E. coli).Here, we develop a mechanistic model of a self-fabricating cell, validated across diverse growth conditions. Through resource balance analysis (RBA), we explore the variation in maximum growth rate with ribosome composition, assuming constant kinetic parameters.Our model highlights the importance of RNA instability. If we neglect it, RNA synthesis is always cheaper than protein synthesis, leading to an RNA-only ribosome at maximum growth rate. Upon accounting for RNA turnover, we find that a mixed ribosome composed of RNA and proteins maximizes growth rate. To account for RNA turnover, we explore two scenarios regarding the activity of RNases. In (a) degradation is proportional to RNA content. In (b) ribosomal proteins cooperatively mitigate RNA instability by protecting it from misfolding and subsequent degradation. In both cases, higher protein content elevates protein synthesis costs and simultaneously lowers RNA turnover expenses, resulting in mixed RNA-protein ribosomes. Only scenario (b) aligns qualitatively with experimental data across varied growth conditions.Our research provides fresh insights into ribosome biogenesis and evolution, paving the way for understanding protein-rich ribosomes in archaea and mitochondria.

Optimizing VLP production in gene therapy: Opportunities and challenges for in silico modeling.

Over the past decades, virus-like particle (VLP)-based gene therapy (GT) evolved as a promising approach to cure inherited diseases or cancer. Tremendous costs due to inefficient production processes remain one of the key challenges despite considerable efforts to improve titers. This review aims to link genome-scale metabolic models (GSMMs) to cell lines used for VLP synthesis for the first time. We summarize recent advances and challenges of GSMMs for Chinese hamster ovary (CHO) cells and provide an overview of potential cell lines used in GT. Although GSMMs in CHO cells led to significant improvements in growth rates and recombinant protein (RP)-production, no GSMM has been established for VLP production so far. To facilitate the generation of GSMM for these cell lines we further provide an overview of existing omics data and the highest production titers so far reported.

Towards rational glyco-engineering in CHO: from data to predictive models.

Metabolic modelling strives to develop modelling approaches that are robust and highly predictive. To achieve this, various modelling designs, including hybrid models, and parameter estimation methods that define the type and number of parameters used in the model, are adapted. Accurate input data play an important role so that the selection of experimental methods that provide input data of the required precision with low measurement errors is crucial. For the biopharmaceutically relevant protein glycosylation, the most prominent available models are kinetic models which are able to capture the dynamic nature of protein N-glycosylation. In this review we focus on how to choose the most suitable model for a specific research question, as well as on parameters and considerations to take into account before planning relevant experiments.

Error propagation in constraint-based modeling of Chinese hamster ovary cells.

Chinese hamster ovary (CHO) cells are the most popular mammalian cell factories for the production of glycosylated biopharmaceuticals. To further increase titer and productivity and ensure product quality, rational system-level engineering strategies based on constraint-based metabolic modeling, such as flux balance analysis (FBA), have gained strong interest. However, the quality of FBA predictions depends on the accuracy of the experimental input data, especially on the exchange rates of extracellular metabolites. Yet, it is not standard practice to devote sufficient attention to the accurate determination of these rates. In this work, we investigated to what degree the sampling frequency during a batch culture and the measurement errors of metabolite concentrations influence the accuracy of the calculated exchange rates and further, how this error then propagates into FBA predictions of growth rates. We determined that accurate measurements of essential amino acids with low uptake rates are crucial for the accuracy of FBA predictions, followed by a sufficient number of analyzed time points. We observed that the measured difference in growth rates of two cell lines can only be reliably predicted when both high measurement accuracy and sampling frequency are ensured.

Robust Analytical Methods for the Accurate Quantification of the Total Biomass Composition of Mammalian Cells.

Biomass composition is an important input for genome-scale metabolic models and has a big impact on their predictive capabilities. However, researchers often rely on generic data for biomass composition, e.g. collected from similar organisms. This leads to inaccurate predictions, because biomass composition varies between different cell lines, conditions, and growth phases. In this chapter we present protocols for the determination of the biomass composition of Chinese Hamster Ovary (CHO) cells. These methods can easily be adapted to other types of mammalian cells. The protocols include the quantification of cell dry mass and of the main biomass components, namely protein, lipid, DNA, RNA, and carbohydrates. Cell dry mass is determined gravimetrically by weighing a defined number of cells. Amino acid composition and protein content are measured by gas chromatography mass spectrometry. Lipids are quantified by shotgun mass spectrometry, which provides quantities for the different lipid classes and also the distribution of fatty acids. RNA is purified and then quantified spectrophotometrically. The methods for DNA and carbohydrates are simple fluorometric and colorimetric assays adapted to a 96-well plate format. To ensure quantitative results, internal standards or spike-in controls are used in all methods, e.g. to account for possible matrix effects or loss of material. Finally, the last section provides a guide on how to convert the measured data into biomass equations, which can then be integrated into a metabolic model.

Modulation of heterologous expression from PBAD promoter in Escherichia coli production strains.

Promoter PBAD is frequently used for heterologous gene expression due to several advantages, such as moderately high expression levels, induction by an inexpensive and non-toxic monosaccharide L-arabinose and tight regulation of transcription, which is particularly important for expression of toxic proteins. A drawback of this promoter is all-or-none induction that occurs at subsaturating inducer concentrations. Although the overall expression level of the cell culture seems to correlate with increasing arabinose concentrations, the population is a mixture of induced and uninduced cells and with increasing arabinose concentrations, only the fraction of induced cells increases. This phenomenon is caused by autocatalytic gene expression - the expression of the arabinose transporter AraE is induced by the transported molecule. In this work the promoter PE, controlling the expression of araE, was exchanged for the stronger PBAD promoter in two Escherichia coli strains commonly used for heterologous protein production. This modification should increase a basal number of arabinose transporters in the cell wall and reduce the threshold concentration required for induction and thus reduce heterogeneity of cell population. Heterogeneity and level of expression in individual cells were analysed by flow cytometry using gfp as a reporter gene. In the strain BL21ai, the promoter exchange increased the number of induced cells at subsaturating arabinose concentrations as well as a yield of protein at saturating inducer concentration. In contrast, the modification did not improve these characteristics in RV308ai. In both strains it was possible to modulate the expression level in induced cells 3-6-fold even at subsaturating arabinose concentrations.