Mimicking CHO large-scale effects in the single multicompartment bioreactor: A new approach to access scale-up behavior.

During the scale-up of biopharmaceutical production processes, insufficiently predictable performance losses may occur alongside gradients and heterogeneities. To overcome such performance losses, tools are required to explain, predict, and ultimately prohibit inconsistencies between laboratory and commercial scale. In this work, we performed CHO fed-batch cultivations in the single multicompartment bioreactor (SMCB), a new scale-down reactor system that offers new access to study large-scale heterogeneities in mammalian cell cultures. At volumetric power inputs of 20.4-1.5 W m-3, large-scale characteristics like long mixing times and dissolved oxygen (DO) heterogeneities were mimicked in the SMCB. Compared to a reference bioreactor (REFB) set-up, the conditions in the SMCB provoked an increase in lactate accumulation of up to 87%, an increased glucose uptake, and reduced viable cell concentrations in the stationary phase. All are characteristic for large-scale performance. The unique possibility to distinguish between the effects of changing power inputs and observed heterogeneities provided new insights into the potential reasons for altered product quality attributes. Apparently, the degree of galactosylation in the evaluated glycan patterns changed primarily due to the different power inputs rather than the provoked heterogeneities. The SMCB system could serve as a potent tool to provide new insights into scale-up behavior and to predict cell line-specific drawbacks at an early stage of process development.

Synthetic co-culture in an interconnected two-compartment bioreactor system: violacein production with recombinant E. coli strains.

The concept of modular synthetic co-cultures holds considerable potential for biomanufacturing, primarily to reduce the metabolic burden of individual strains by sharing tasks among consortium members. However, current consortia often show unilateral relationships solely, without stabilizing feedback control mechanisms, and are grown in a shared cultivation setting. Such 'one pot' approaches hardly install optimum growth and production conditions for the individual partners. Hence, novel mutualistic, self-coordinating consortia are needed that are cultured under optimal growth and production conditions for each member. The heterologous production of the antibiotic violacein (VIO) in the mutually interacting E. coli-E. coli consortium serves as an example of this new principle. Interdependencies for growth control were implemented via auxotrophies for L-tryptophan and anthranilate (ANT) that were satisfied by the respective partner. Furthermore, VIO production was installed in the ANT auxotrophic strain. VIO production, however, requires low temperatures of 20-30 °C which conflicts with the optimum growth temperature of E. coli at 37 °C. Consequently, a two-compartment, two-temperature level setup was used, retaining the mutual interaction of the cells via the filter membrane-based exchange of medium. This configuration also provided the flexibility to perform individualized batch and fed-batch strategies for each co-culture member. We achieved maximum biomass-specific productivities of around 6 mg (g h)-1 at 25 °C which holds great promise for future applications.

Towards valorization of pectin-rich agro-industrial residues: Engineering of Saccharomyces cerevisiae for co-fermentation of d-galacturonic acid and glycerol.

Pectin-rich plant biomass residues represent underutilized feedstocks for industrial biotechnology. The conversion of the oxidized monomer d-galacturonic acid (d-GalUA) to highly reduced fermentation products such as alcohols is impossible due to the lack of electrons. The reduced compound glycerol has therefore been considered an optimal co-substrate, and a cell factory able to efficiently co-ferment these two carbon sources is in demand. Here, we inserted the fungal d-GalUA pathway in a strain of the yeast S. cerevisiae previously equipped with an NAD-dependent glycerol catabolic pathway. The constructed strain was able to consume d-GalUA with the highest reported maximum specific rate of 0.23 g gCDW-1 h-1 in synthetic minimal medium when glycerol was added. By means of a 13C isotope-labelling analysis, carbon from both substrates was shown to end up in pyruvate. The study delivers the proof of concept for a co-fermentation of the two 'respiratory' carbon sources to ethanol and demonstrates a fast and complete consumption of d-GalUA in crude sugar beet pulp hydrolysate under aerobic conditions. The future challenge will be to achieve co-fermentation under industrial, quasi-anaerobic conditions.

Engineering of a robust Escherichia coli chassis and exploitation for large-scale production processes.

In large-scale bioprocesses microbes are exposed to heterogeneous substrate availability reducing the overall process performance. A series of deletion strains was constructed from E. coli MG1655 aiming for a robust phenotype in heterogeneous fermentations with transient starvation. Deletion targets were hand-picked based on a list of genes derived from previous large-scale simulation runs. Each gene deletion was conducted on the premise of strict neutrality towards growth parameters in glucose minimal medium. The final strain of the series, named E. coli RM214, was cultivated continuously in an STR-PFR (stirred tank reactor - plug flow reactor) scale-down reactor. The scale-down reactor system simulated repeated passages through a glucose starvation zone. When exposed to nutrient gradients, E. coli RM214 had a significantly lower maintenance coefficient than E. coli MG1655 (Δms = 0.038 gGlucose/gCDW/h, p < 0.05). In an exemplary protein production scenario E. coli RM214 remained significantly more productive than E. coli MG1655 reaching 44% higher eGFP yield after 28 h of STR-PFR cultivation. This study developed E. coli RM214 as a robust chassis strain and demonstrated the feasibility of engineering microbial hosts for large-scale applications.

Data-driven in silico prediction of regulation heterogeneity and ATP demands of Escherichia coli in large-scale bioreactors.

Escherichia coli exposed to industrial-scale heterogeneous mixing conditions respond to external stress by initiating short-term metabolic and long-term strategic transcriptional programs. In native habitats, long-term strategies allow survival in severe stress but are of limited use in large bioreactors, where microenvironmental conditions may change right after said programs are started. Related on/off switching of genes causes additional ATP burden that may reduce the cellular capacity for producing the desired product. Here, we present an agent-based data-driven model linked to computational fluid dynamics, finally allowing to predict additional ATP needs of Escherichia coli K12 W3110 exposed to realistic large-scale bioreactor conditions. The complex model describes transcriptional up- and downregulation dynamics of about 600 genes starting from subminute range covering 28 h. The data-based approach was extracted from comprehensive scale-down experiments. Simulating mixing and mass transfer conditions in a 54 m3 stirred bioreactor, 120,000 E. coli cells were tracked while fluctuating between different zones of glucose availability. It was found that cellular ATP demands rise between 30% and 45% of growth decoupled maintenance needs, which may limit the production of ATP-intensive product formation accordingly. Furthermore, spatial analysis of individual cell transcriptional patterns reveal very heterogeneous gene amplifications with hot spots of 50%-80% messenger RNA upregulation in the upper region of the bioreactor. The phenomenon reflects the time-delayed regulatory response of the cells that propagate through the stirred tank. After 4.2 h, cells adapt to environmental changes but still have to bear an additional 6% ATP demand.

Balancing glucose and oxygen uptake rates to enable high amorpha-4,11-diene production in Escherichia coli via the methylerythritol phosphate pathway.

Amorpha-4,11-diene (AMD4,11) is a precursor to artemisinin, a potent antimalarial drug that is traditionally extracted from the shrubs of Artemisia annua. Despite significant prior efforts to produce artemisinin and its precursors through biotechnology, there remains a dire need for more efficient biosynthetic routes for its production. Here, we describe the optimization of key process conditions for an Escherichia coli strain producing AMD4,11 via the native methylerythritol phosphate (MEP) pathway. By studying the interplay between glucose uptake rates and oxygen demand, we were able to identify optimal conditions for increasing carbon flux through the MEP pathway by manipulating the availability of NADPH required for terpenoid production. Installation of an optimal qO2 /qglucose led to a 6.7-fold increase in product titers and a 6.5-fold increase in carbon yield.

Pseudomonas putida KT2440 endures temporary oxygen limitations.

The obligate aerobic nature of Pseudomonas putida, one of the most prominent whole-cell biocatalysts emerging for industrial bioprocesses, questions its ability to be cultivated in large-scale bioreactors, which exhibit zones of low dissolved oxygen tension. P. putida KT2440 was repeatedly subjected to temporary oxygen limitations in scale-down approaches to assess the effect on growth and an exemplary production of rhamnolipids. At those conditions, the growth and production of P. putida KT2440 were decelerated compared to well-aerated reference cultivations, but remarkably, final biomass and rhamnolipid titers were similar. The robust growth behavior was confirmed across different cultivation systems, media compositions, and laboratories, even when P. putida KT2440 was repeatedly exposed to dual carbon and oxygen starvation. Quantification of the nucleotides ATP, ADP, and AMP revealed a decrease of intracellular ATP concentrations with increasing duration of oxygen starvation, which can, however, be restored when re-supplied with oxygen. Only small changes in the proteome were detected when cells encountered oscillations in dissolved oxygen tensions. Concluding, P. putida KT2440 appears to be able to cope with repeated oxygen limitations as they occur in large-scale bioreactors, affirming its outstanding suitability as a whole-cell biocatalyst for industrial-scale bioprocesses.

Compartment-specific metabolome labeling enables the identification of subcellular fluxes that may serve as promising metabolic engineering targets in CHO cells.

13C labeling data are used to calculate quantitative intracellular flux patterns reflecting in vivo conditions. Given that approaches for compartment-specific metabolomics exist, the benefits they offer compared to conventional non-compartmented 13C flux studies remain to be determined. Using compartment-specific labeling information of IgG1-producing Chinese hamster ovary cells, this study investigated differences of flux patterns exploiting and ignoring metabolic labeling data of cytosol and mitochondria. Although cellular analysis provided good estimates for the majority of intracellular fluxes, half of the mitochondrial transporters, and NADH and ATP balances, severe differences were found for some reactions. Accurate flux estimations of almost all iso-enzymes heavily depended on the sub-cellular labeling information. Furthermore, key discrepancies were found for the mitochondrial carriers vAGC1 (Aspartate/Glutamate antiporter), vDIC (Malate/H+ symporter), and vOGC (α-ketoglutarate/malate antiporter). Special emphasis is given to the flux of cytosolic malic enzyme (vME): it could not be estimated without the compartment-specific malate labeling information. Interesting enough, cytosolic malic enzyme is an important metabolic engineering target for improving cell-specific IgG1 productivity. Hence, compartment-specific 13C labeling analysis serves as prerequisite for related metabolic engineering studies.

Prediction of novel non-coding RNAs relevant for the growth of Pseudomonas putida in a bioreactor.

Pseudomonas putida is a micro-organism with great potential for industry due to its stress-endurance traits and easy manipulation of the metabolism. However, optimization is still required to improve production yields. In the last years, manipulation of bacterial small non-coding RNAs (ncRNAs) has been recognized as an effective tool to improve the production of industrial compounds. So far, very few ncRNAs are annotated in P. putida beyond the generally conserved. In the present study, P. putida was cultivated in a two-compartment scale-down bioreactor that simulates large-scale industrial bioreactors. We performed RNA-Seq of samples collected at distinct locations and time-points to predict novel and potentially important ncRNAs for the adaptation of P. putida to bioreactor stress conditions. Instead of using a purely genomic approach, we have rather identified regions of putative ncRNAs with high expression levels using two different programs (Artemis and sRNA detect). Only the regions identified with both approaches were considered for further analysis and, in total, 725 novel ncRNAs were predicted. We also found that their expression was not constant throughout the bioreactor, showing different patterns of expression with time and position. This is the first work focusing on the ncRNAs whose expression is triggered in a bioreactor environment. This information is of great importance for industry, since it provides possible targets to engineer more effective P. putida strains for large-scale production.

Simulated oxygen and glucose gradients as a prerequisite for predicting industrial scale performance a priori.

Transferring bioprocesses from lab to industrial scale without loss of performance is key for the successful implementation of novel production approaches. Because mixing and mass transfer is usually hampered in large scale, cells experience heterogeneities eventually causing deteriorated yields, that is, reduced titers, productivities, and sugar-to-product conversions. Accordingly, reliable and easy-to-implement tools for a priori prediction of large-scale performance based on dry and wet-lab tests are heavily needed. This study makes use of computational fluid dynamic simulations of a multiphase multi-impeller stirred tank in pilot scale. So-called lifelines, records of 120,000 Corynebacterium glutamicum cells experiencing fluctuating environmental conditions, were identified and used to properly design wet-lab scale-down (SD) devices. Physical parameters such as power input, gas hold up, kLa , and mixing time showed good agreement with experimental measurements. Analyzing the late fed-batch cultivation revealed that the complex double gradient of glucose and oxygen can be translated into a wet-lab SD setup with only few compartments. Most remarkably, the comparison of different mesh sizes outlined that even the coarsest approach with a mesh density of 1.12×105#/m3 was sufficient to properly predict physical and biological readouts. Accordingly, the approach offers the potential for the thorough analysis of realistic industrial case scenarios.

S-adenosylmethionine and methylthioadenosine boost cellular productivities of antibody forming Chinese hamster ovary cells.

The improvement of cell specific productivities for the formation of therapeutic proteins is an important step towards intensified production processes. Among others, the induction of the desired production phenotype via proper media additives is a feasible solution provided that said compounds adequately trigger metabolic and regulatory programs inside the cells. In this study, S-(5'-adenosyl)- l-methionine (SAM) and 5'-deoxy-5'-(methylthio)adenosine (MTA) were found to stimulate cell specific productivities up to approx. 50% while keeping viable cell densities transiently high and partially arresting the cell cycle in an anti-IL-8-producing CHO-DP12 cell line. Noteworthy, MTA turned out to be the chemical degradation product of the methyl group donor SAM and is consumed by the cells.

From nutritional wealth to autophagy: In vivo metabolic dynamics in the cytosol, mitochondrion and shuttles of IgG producing CHO cells.

To fulfil the optimization needs of current biopharmaceutical processes the knowledge how to improve cell specific productivities is of outmost importance. This requires a detailed understanding of cellular metabolism on a subcellular level inside compartments such as cytosol and mitochondrion. Using IgG1 producing Chinese hamster ovary (CHO) cells, a pioneering protocol for compartment-specific metabolome analysis was applied. Various production-like growth conditions ranging from ample glucose and amino acid supply via moderate to severe nitrogen limitation were investigated in batch cultures. The combined application of quantitative metabolite pool analysis, 13C tracer studies and non-stationary flux calculations revealed that Pyr/H+ symport (MPC1/2) bore the bulk of the mitochondrial transport under ample nutrient supply. Glutamine limitation induced the concerted adaptation of the bidirectional Mal/aKG (OGC) and the Mal/HPO42- antiporter (DIC), even installing completely reversed shuttle fluxes. As a result, NADPH and ATP formation were adjusted to cellular needs unraveling the key role of cytosolic malic enzyme for NADPH production. Highest cell specific IgG1 productivities were closely correlated to a strong mitochondrial malate export according to the anabolic demands. The requirement to install proper NADPH supply for optimizing the production of monoclonal antibodies is clearly outlined. Interestingly, it was observed that mitochondrial citric acid cycle activity was always maintained enabling constant cytosolic adenylate energy charges at physiological levels, even under autophagy conditions.

Exploiting Hydrogenophaga pseudoflava for aerobic syngas-based production of chemicals.

Gasification is a suitable technology to generate energy-rich synthesis gas (syngas) from biomass or waste streams, which can be utilized in bacterial fermentation processes for the production of chemicals and fuels. Established microbial processes currently rely on acetogenic bacteria which perform an energetically inefficient anaerobic CO oxidation and acetogenesis potentially hampering the biosynthesis of complex and ATP-intensive products. Since aerobic oxidation of CO is energetically more favorable, we exploit in this study the Gram-negative ß-proteobacterium Hydrogenophaga pseudoflava DSM1084 as novel host for the production of chemicals from syngas. We sequenced and annotated the genome of H. pseudoflava and established a genetic engineering toolbox, which allows markerless chromosomal modification via the pk19mobsacB system and heterologous gene expression on pBBRMCS2-based plasmids. The toolbox was extended by identifying strong endogenous promotors such as PgapA2 which proved to yield high expression under heterotrophic and autotrophic conditions. H. pseudoflava showed relatively fast heterotrophic growth in complex and minimal medium with sugars and organic acids which allows convenient handling in lab routines. In autotrophic bioreactor cultivations with syngas, H. pseudoflava exhibited a growth rate of 0.06 h-1 and biomass specific uptakes rates of 14.2 ±â€¯0.3 mmol H2 gCDW-1 h-1, 73.9 ±â€¯1.8 mmol CO gCDW-1 h-1, and 31.4 ±â€¯0.3 mmol O2 gCDW-1 h-1. As proof of concept, we engineered the carboxydotrophic bacterium for the aerobic production of the C15 sesquiterpene (E)-α-bisabolene from the C1 carbon source syngas by heterologous expression of the (E)-α-bisabolene synthase gene agBIS. The resulting strain H. pseudoflava (pOCEx1:agBIS) produced 59 ±â€¯8 µg (E)-α-bisabolene L-1 with a volumetric productivity Qp of 1.2 ±â€¯0.2 µg L-1 h-1 and a biomass-specific productivity qp of 13.1 ±â€¯0.6 µg gCDW-1 h-1. The intrinsic properties and the genetic repertoire of H. pseudoflava make this carboxydotrophic bacterium a promising candidate for future aerobic production processes to synthesize more complex or ATP-intensive chemicals from syngas.

Perfusion cultures require optimum respiratory ATP supply to maximize cell-specific and volumetric productivities.

Perfusion processes are an emerging alternative to common fed-batch processes in the growing biopharmaceutical industry. However, the challenge of maintaining high cell-specific productivities remains. In this study, glucose limitation was applied to two perfusion steady states and compared with a third steady state without any detectable limitation. The metabolic phenotype was enhanced under glucose limitation with a decrease of 30% in glucose uptake and 75% in lactate formation. Cell-specific productivities were substantially improved by 50%. Remarkably, the productivities showed a strong correlation to respiratory adenosine triphosphate (ATP) supply. As less reduced nicotinamide adenine dinucleotide (NADH) remained in the cytosol, the ATP generation from oxidative phosphorylation was increased by almost 30%. Consequently, the efficiency of carbon metabolism and the resulting respiratory ATP supply was crucial for maintaining the highly productive cellular state. This study highlights that glucose limitation can be used for process intensification in perfusion cultures as ATP generation via respiration is significantly increased, leading to elevated productivities.

Protein production in Escherichia coli is guided by the trade-off between intracellular substrate availability and energy cost.

BACKGROUND: In vivo protein formation is a crucial part of cellular life. The process needs to adapt to growth conditions and is exploited for the production of technical and pharmaceutical proteins in microbes such as Escherichia coli. Accordingly, the elucidation of basic regulatory mechanisms controlling the in vivo translation machinery is of primary interest, not only to improve heterologous protein production but also to elucidate fundamental regulation regimens of cellular growth. RESULTS: The current modeling analysis elucidates the impact of diffusion for the stochastic supply of crucial substrates such as the elongation factor EFTu, and tRNA species, all regarded as key elements for ensuring optimum transcriptional elongation. Together with the consideration of cellular ribosome numbers, their impact on the proper functioning of the translation machinery was investigated under different in vivo and in vitro conditions and utilizing the formation of non-native GFP and native EFTu as target proteins. The results show that translational elongation was diffusion limited. However, this effect was much more pronounced for the translation of non-native proteins than for the formation of codon-optimized native proteins. CONCLUSIONS: Cellular ATP requirements constrain the options of improving protein production. In the case of non-native protein sequences, an optimized tRNA supply may be the most economical solution, as cells necessarily have to invest in ATP-costly ribosome synthesis to boost translation and increase growth rates.

Metabolic engineering to guide evolution - Creating a novel mode for L-valine production with Corynebacterium glutamicum.

Evolutionary approaches are often undirected and mutagen-based yielding numerous mutations, which need elaborate screenings to identify relevant targets. We here apply Metabolic engineering to Guide Evolution (MGE), an evolutionary approach evolving and identifying new targets to improve microbial producer strains. MGE is based on the idea to impair the cell's metabolism by metabolic engineering, thereby generating guided evolutionary pressure. It consists of three distinct phases: (i) metabolic engineering to create the evolutionary pressure on the applied strain followed by (ii) a cultivation phase with growth as straightforward screening indicator for the evolutionary event, and (iii) comparative whole genome sequencing (WGS), to identify mutations in the evolved strains, which are eventually re-engineered for verification. Applying MGE, we evolved the PEP and pyruvate carboxylase-deficient strain C. glutamicum Δppc Δpyc to grow on glucose as substrate with rates up to 0.31 ±â€¯0.02 h-1 which corresponds to 80% of the growth rate of the wildtype strain. The intersection of the mutations identified by WGS revealed isocitrate dehydrogenase (ICD) as consistent target in three independently evolved mutants. Upon re-engineering in C. glutamicum Δppc Δpyc, the identified mutations led to diminished ICD activities and activated the glyoxylate shunt replenishing oxaloacetate required for growth. Intracellular relative quantitative metabolome analysis showed that the pools of citrate, isocitrate, cis-aconitate, and L-valine were significantly higher compared to the WT control. As an alternative to existing L-valine producer strains based on inactivated or attenuated pyruvate dehydrogenase complex, we finally engineered the PEP and pyruvate carboxylase-deficient C. glutamicum strains with identified ICD mutations for L-valine production by overexpression of the L-valine biosynthesis genes. Among them, C. glutamicum Δppc Δpyc ICDG407S (pJC4ilvBNCE) produced up to 8.9 ±â€¯0.4 g L-valine L-1, with a product yield of 0.22 ±â€¯0.01 g L-valine per g glucose.

Escherichia coli HGT: Engineered for high glucose throughput even under slowly growing or resting conditions.

Aerobic production-scale processes are constrained by the technical limitations of maximum oxygen transfer and heat removal. Consequently, microbial activity is often controlled via limited nutrient feeding to maintain it within technical operability. Here, we present an alternative approach based on a newly engineered Escherichia coli strain. This E. coli HGT (high glucose throughput) strain was engineered by modulating the stringent response regulation program and decreasing the activity of pyruvate dehydrogenase. The strain offers about three-fold higher rates of cell-specific glucose uptake under nitrogen-limitation (0.6gGlc gCDW-1h-1) compared to that of wild type, with a maximum glucose uptake rate of about 1.8gGlc gCDW-1h-1 already at a 0.3h-1 specific growth rate. The surplus of imported glucose is almost completely available via pyruvate and is used to fuel pyruvate and lactate formation. Thus, E. coli HGT represents a novel chassis as a host for pyruvate-derived products.

High Substrate Uptake Rates Empower Vibrio natriegens as Production Host for Industrial Biotechnology.

The productivity of industrial fermentation processes is essentially limited by the biomass-specific substrate consumption rate (qS ) of the applied microbial production system. Since qS depends on the growth rate (µ), we highlight the potential of the fastest-growing nonpathogenic bacterium, Vibrio natriegens, as a novel candidate for future biotechnological processes. V. natriegens grows rapidly in BHIN complex medium with a µ of up to 4.43 h-1 (doubling time of 9.4 min) as well as in minimal medium supplemented with various industrially relevant substrates. Bioreactor cultivations in minimal medium with glucose showed that V. natriegens possesses an exceptionally high qS under aerobic (3.90 ± 0.08 g g-1 h-1) and anaerobic (7.81 ± 0.71 g g-1 h-1) conditions. Fermentations with resting cells of genetically engineered V. natriegens under anaerobic conditions yielded an overall volumetric productivity of 0.56 ± 0.10 g alanine liter-1 min-1 (i.e., 34 g liter-1 h-1). These inherent properties render V. natriegens a promising new microbial platform for future industrial fermentation processes operating with high productivity.IMPORTANCE Low conversion rates are one major challenge to realizing microbial fermentation processes for the production of commodities operating competitively with existing petrochemical approaches. For this reason, we screened for a novel platform organism possessing characteristics superior to those of traditionally employed microbial systems. We identified the fast-growing V. natriegens, which exhibits a versatile metabolism and shows striking growth and conversion rates, as a solid candidate to reach outstanding productivities. Due to these inherent characteristics, V. natriegens can speed up common laboratory routines, is suitable for already existing production procedures, and forms an excellent foundation for engineering next-generation bioprocesses.

Dynamic modeling reveals a three-step response of Saccharomyces cerevisiae to high CO2 levels accompanied by increasing ATP demands.

Saccharomyces cerevisiae is often applied in large-scale bioreactors where gradients of dissolved CO2 exist. Under high CO2 pressure, the dissolved gas enters the microbe, causing multifold intracellular responses such as decrease of pH, increase of HCO3- and changes of ion balance. Effects of varying CO2 concentrations are multifold, hard to scale and hardly investigated. Hence, the multi-level response to CO2 shifts was summarized in a predicting ODE model with mass action kinetics, balancing electrochemical charges in steady-state growth conditions. Compared to experimental observations, the simulated dynamics of ion concentrations were found to be consistent. During CO2 shifts, the model predicts the initial depolarization of the membrane potential, the temporal pH drop and the activation of countermeasures such as Pma1-mediated H+ export and Trk1,2-mediated K+ import. In conclusion, extracellular cation concentrations and the cellular pH regulation are critical factors that determine physiology and cellular energy management. Consequently, pressure-induced CO2 gradients cause peaks of ATP demand which may occur in cells circulating in large-scale industrial bioreactors.

Engineering E. coli for large-scale production - Strategies considering ATP expenses and transcriptional responses.

Microbial producers such as Escherichia coli are evolutionarily trained to adapt to changing substrate availabilities. Being exposed to large-scale production conditions, their complex, multilayered regulatory programs are frequently activated because they face changing substrate supply due to limited mixing. Here, we show that E. coli can adopt both short- and long-term strategies to withstand these stress conditions. Experiments in which glucose availability was changed over a short time scale were performed in a two-compartment bioreactor system. Quick metabolic responses were observed during the first 30s of glucose shortage, and after 70s, fundamental transcriptional programs were initiated. Since cells are fluctuating under simulated large-scale conditions, this scenario represents a continuous on/off switching of about 600 genes. Furthermore, the resulting ATP maintenance demands were increased by about 40-50%, allowing us to conclude that hyper-producing strains could become ATP-limited under large-scale production conditions. Based on the observed transcriptional patterns, we identified a number of candidate gene deletions that may reduce unwanted ATP losses. In summary, we present a theoretical framework that provides biological targets that could be used to engineer novel E. coli strains such that large-scale performance equals laboratory-scale expectations.