Design of growth‐dependent biosensors based on destabilized GFP for the detection of physiological behavior of Escherichia coli in heterogeneous bioreactors

In this work, we present the design and characterization of Green Fluorescent Protein (GFP)‐based reporter systems designed to describe cellular activity in “complex,” heterogeneous bioreactors. The reporter systems consist of Escherichia coli strains carrying growth dependent promoters fused to genes expressing stable and unstable variants of GFP, respectively. The response of Escherichia coli cells to transient exposure to glucose was studied in a two‐compartment scale down bioreactor (SDR) consisting of a well‐stirred tank reactor (STR) connected to a plug‐flow reactor (PFR). Such a SDR system is employed to mimic the situation of high glucose concentration and oxygen limitation that often encountered in large‐scale, fed‐batch bioreactors and the response of E. coli was simulated by continuously pumping microbial cells from STR to the PFR. We found that repeated addition of concentrated glucose pulses with varied frequency at the entrance of the PFR had consequences on strain physiological behavior. The GFP expressions were significantly marked after 10 h of cultivation in STR (control reactor) and SDR, whereas, growth rates were rather similar. Additional experiments in chemostat with programmed glucose perturbation suggested that the activities of the promoters were linked with the substrate limitation signal. Taken together with immunoblot analysis, we suppose protein leakage is responsible for the overexpression of fis and the related promoters, such as rrnB in this case study, but additional works are required in order to confirm this relationship. This investigation is useful for a better understanding of the fast dynamic phenomena occurring in heterogeneous large‐scale bioreactors. © 2013 American Institute of Chemical Engineers Biotechnol. Prog., 29: 553–563, 2013     

Fast dynamic response of the fermentative metabolism of Escherichia coli to aerobic and anaerobic glucose pulses

The response of Escherichia coli cells to transient exposure (step increase) in substrate concentration and anaerobiosis leading to mixed‐acid fermentation metabolism was studied in a two‐compartment bioreactor system consisting of a stirred tank reactor (STR) connected to a mini‐plug‐flow reactor (PFR: BioScope, 3.5 mL volume). Such a system can mimic the situation often encountered in large‐scale, fed‐batch bioreactors. The STR represented the zones of a large‐scale bioreactor that are far from the point of substrate addition and that can be considered as glucose limited, whereas the PFR simulated the region close to the point of substrate addition, where glucose concentration is much higher than in the rest of the bioreactor. In addition, oxygen‐poor and glucose‐rich regions can occur in large‐scale bioreactors. The response of E. coli to these large‐scale conditions was simulated by continuously pumping E. coli cells from a well stirred, glucose limited, aerated chemostat (D = 0.1 h^?1) into the mini‐PFR. A glucose pulse was added at the entrance of the PFR. In the PFR, a total of 11 samples were taken in a time frame of 92 s. In one case aerobicity in the PFR was maintained in order to evaluate the effects of glucose overflow independently of oxygen limitation. Accumulation of acetate and formate was detected after E. coli cells had been exposed for only 2 s to the glucose‐rich (aerobic) region in the PFR. In the other case, the glucose pulse was also combined with anaerobiosis in the PFR. Glucose overflow combined with anaerobiosis caused the accumulation of formate, acetate, lactate, ethanol, and succinate, which were also detected as soon as 2 s after of exposure of E. coli cells to the glucose and O₂ gradients. This approach (STR‐mini‐PFR) is useful for a better understanding of the fast dynamic phenomena occurring in large‐scale bioreactors and for the design of modified strains with an improved behavior under large‐scale conditions. Biotechnol. Bioeng. 2009; 104: 1153–1161. © 2009 Wiley Periodicals, Inc.     

Application of high‐throughput mini‐bioreactor system for systematic scale‐down modeling,process characterization,and control strategy development

High‐throughput systems and processes have typically been targeted for process development and optimization in the bioprocessing industry. For process characterization, bench scale bioreactors have been the system of choice. Due to the need for performing different process conditions for multiple process parameters, the process characterization studies typically span several months and are considered time and resource intensive. In this study, we have shown the application of a high‐throughput mini‐bioreactor system viz. the Advanced Microscale Bioreactor (ambr15^TM), to perform process characterization in less than a month and develop an input control strategy. As a pre‐requisite to process characterization, a scale‐down model was first developed in the ambr system (15 mL) using statistical multivariate analysis techniques that showed comparability with both manufacturing scale (15,000 L) and bench scale (5 L). Volumetric sparge rates were matched between ambr and manufacturing scale, and the ambr process matched the pCO₂ profiles as well as several other process and product quality parameters. The scale‐down model was used to perform the process characterization DoE study and product quality results were generated. Upon comparison with DoE data from the bench scale bioreactors, similar effects of process parameters on process yield and product quality were identified between the two systems. We used the ambr data for setting action limits for the critical controlled parameters (CCPs), which were comparable to those from bench scale bioreactor data. In other words, the current work shows that the ambr15^TM system is capable of replacing the bench scale bioreactor system for routine process development and process characterization. © 2015 American Institute of Chemical Engineers Biotechnol. Prog., 31:1623–1632, 2015     

Application of multivariate analysis and mass transfer principles for refinement of a 3‐L bioreactor scale‐down model—when shake flasks mimic 15,000‐L bioreactors better

Characterization of manufacturing processes is key to understanding the effects of process parameters on process performance and product quality. These studies are generally conducted using small‐scale model systems. Because of the importance of the results derived from these studies, the small‐scale model should be predictive of large scale. Typically, small‐scale bioreactors, which are considered superior to shake flasks in simulating large‐scale bioreactors, are used as the scale‐down models for characterizing mammalian cell culture processes. In this article, we describe a case study where a cell culture unit operation in bioreactors using one‐sided pH control and their satellites (small‐scale runs conducted using the same post‐inoculation cultures and nutrient feeds) in 3‐L bioreactors and shake flasks indicated that shake flasks mimicked the large‐scale performance better than 3‐L bioreactors. We detail here how multivariate analysis was used to make the pertinent assessment and to generate the hypothesis for refining the existing 3‐L scale‐down model. Relevant statistical techniques such as principal component analysis, partial least square, orthogonal partial least square, and discriminant analysis were used to identify the outliers and to determine the discriminatory variables responsible for performance differences at different scales. The resulting analysis, in combination with mass transfer principles, led to the hypothesis that observed similarities between 15,000‐L and shake flask runs, and differences between 15,000‐L and 3‐L runs, were due to pCO₂ and pH values. This hypothesis was confirmed by changing the aeration strategy at 3‐L scale. By reducing the initial sparge rate in 3‐L bioreactor, process performance and product quality data moved closer to that of large scale. © 2015 American Institute of Chemical Engineers Biotechnol. Prog., 31:1370–1380, 2015     

Optical sensor enabled rocking T-flasks as novel upstream bioprocessing tools

During the past decade, novel disposable cell culture vessels (generally referred to as Process Scouting Devices or PSDs) have become increasingly popular for laboratory scale studies and seed culture generation. However, the lack of engineering characterization and online monitoring tools for PSDs makes it difficult to elucidate their oxygen transfer capabilities. In this study, a mass transfer characterization (k_La) of sensor enabled static and rocking T‐flasks is presented and compared with other non‐instrumented PSDs such as CultiFlask 50®, spinner flasks, and SuperSpinner D 1000®. We have also developed a mass transfer empirical correlation that accounts for the contribution of convection and diffusion to the volumetric mass transfer coefficient (k_La) in rocking T‐flasks. We also carried out a scale‐down study at matched k_La between a rocking T75‐flask and a 10 L (2 L filling volume) wave bioreactor (Cultibag®) and we observed similar DO and pH profiles as well as maximum cell density and protein titer. However, in this scale‐down study, we also observed a negative correlation between cell growth and protein productivity between the rocking T‐flask and the wave bioreactor. We hypothesize that this negative correlation can be due to hydrodynamic stress difference between the rocking T‐flask and the Cultibag. As both cell culture devices share key similarities such as type of agitation (i.e., rocking), oxygen transfer capabilities (i.e., k_La) and disposability, we argue that rocking T‐flasks can be readily integrated with wave bioreactors, making the transition from research‐scale to manufacturing‐scale a seamless process. Biotechnol. Bioeng. 2012;109: 2295–2305. © 2012 Wiley Periodicals, Inc.     

Software sensor design considering oscillating conditions as present in industrial scale fed‐batch cultivations

Investigations of inhomogeneous dynamics in industrial‐scale bioreactors can be realized in laboratory simulators. Such studies will be improved by on line observation of the growth of microorganisms and their product synthesis at oscillating substrate availability which represents the conditions in industrial‐scale fed‐batch cultivations. A method for the kinetic monitoring of such processes, supported by on line measurements accessible in industrial practice, is proposed. It consists of a software sensor (SS) system composed of a cascade structure. Process kinetics are simulated in models with a structure including time‐varying yield coefficients. SSs for measured variable kinetics have classical structures. The SS design of unmeasured variables is realized after a linear transformation using a logarithmic function. For these software sensors, a tuning procedure is proposed, at which an arbitrary choice of one tuning parameter value that guarantees stability of the monitoring system allows the calculation of the optimal values of six parameters. The effectiveness of the proposed monitoring approach is demonstrated with experimental data from a glucose‐limited fed‐batch process of Bacillus subtilis in a laboratory two‐compartment scale down reactor which tries to mimic the conditions present in industrial scale nutrient‐limited fed‐batch cultivations. Biotechnol. Bioeng. 2013; 110: 1945–1955. © 2013 Wiley Periodicals, Inc.     

A simple eccentric stirred tank mini‐bioreactor: Mixing characterization and mammalian cell culture experiments

In industrial practice, stirred tank bioreactors are the most common mammalian cell culture platform. However, research and screening protocols at the laboratory scale (i.e., 5–100 mL) rely primarily on Petri dishes, culture bottles, or Erlenmeyer flasks. There is a clear need for simple—easy to assemble, easy to use, easy to clean—cell culture mini‐bioreactors for lab‐scale and/or screening applications. Here, we study the mixing performance and culture adequacy of a 30 mL eccentric stirred tank mini‐bioreactor. A detailed mixing characterization of the proposed bioreactor is presented. Laser induced fluorescence (LIF) experiments and computational fluid dynamics (CFD) computations are used to identify the operational conditions required for adequate mixing. Mammalian cell culture experiments were conducted with two different cell models. The specific growth rate and the maximum cell density of Chinese hamster ovary (CHO) cell cultures grown in the mini‐bioreactor were comparable to those observed for 6‐well culture plates, Erlenmeyer flasks, and 1 L fully instrumented bioreactors. Human hematopoietic stem cells were successfully expanded tenfold in suspension conditions using the eccentric mini‐bioreactor system. Our results demonstrate good mixing performance and suggest the practicality and adequacy of the proposed mini‐bioreactor. Biotechnol. Bioeng. 2013; 110: 1106–1118. © 2012 Wiley Periodicals, Inc.     

Characterization of TAP Ambr 250 disposable bioreactors,as a reliable scale‐down model for biologics process development

Demands for development of biological therapies is rapidly increasing, as is the drive to reduce time to patient. In order to speed up development, the disposable Automated Microscale Bioreactor (Ambr 250) system is increasingly gaining interest due to its advantages, including highly automated control, high throughput capacity, and short turnaround time. Traditional early stage upstream process development conducted in 2 ‐ 5 L bench‐top bioreactors requires high foot‐print, and running cost. The establishment of the Ambr 250 as a scale‐down model leads to many benefits in process development. In this study, a comprehensive characterization of mass transfer coefficient (k_La) in the Ambr 250 was conducted to define optimal operational conditions. Scale‐down approaches, including dimensionless volumetric flow rate (vvm), power per unit volume (P/V) and k_La have been evaluated using different cell lines. This study demonstrates that the Ambr 250 generated comparable profiles of cell growth and protein production, as seen at 5‐L and 1000‐L bioreactor scales, when using k_La as a scale‐down parameter. In addition to mimicking processes at large scales, the suitability of the Ambr 250 as a tool for clone selection, which is traditionally conducted in bench‐top bioreactors, was investigated. Data show that cell growth, productivity, metabolite profiles, and product qualities of material generated using the Ambr 250 were comparable to those from 5‐L bioreactors. Therefore, Ambr 250 can be used for clone selection and process development as a replacement for traditional bench‐top bioreactors minimizing resource utilization during the early stages of development in the biopharmaceutical industry. © 2017 American Institute of Chemical Engineers Biotechnol. Prog., 33:478–489, 2017     

Flux response of glycolysis and storage metabolism during rapid feast/famine conditions in Penicillium chrysogenum using dynamic 13C labeling

The scale‐up of fermentation processes frequently leads to a reduced productivity compared to small‐scale screening experiments. Large‐scale mixing limitations that lead to gradients in substrate and oxygen availability could influence the microorganism performance. Here, the impact of substrate gradients on a penicillin G producing Penicillium chrysogenum cultivation was analyzed using an intermittent glucose feeding regime. The intermittent feeding led to fluctuations in the extracellular glucose concentration between 400 μM down to 6.5 μM at the end of the cycle. The intracellular metabolite concentrations responded strongly and showed up to 100‐fold changes. The intracellular flux changes were estimated on the basis of dynamic ¹³C mass isotopomer measurements during three cycles of feast and famine using a novel hybrid modeling approach. The flux estimations indicated a high turnover of internal and external storage metabolites in P. chrysogenum under feast/famine conditions. The synthesis and degradation of storage requires cellular energy (ATP and UTP) in competition with other cellular functions including product formation. Especially, 38% of the incoming glucose was recycled once in storage metabolism. This result indicated that storage turnover is increased under dynamic cultivation conditions and contributes to the observed decrease in productivity compared to reference steady‐state conditions.     

Evaluation of a set of E. coli reporter strains as physiological tracer for estimating bioreactor hydrodynamic efficiency

A set of different green fluorescent protein (GFP) Escherichia coli reporter strains have been evaluated in mini- and stirred bioreactors operating in fed-batch mode with different degrees of perturbations in order to estimate their potential use as process-related stress biosensor. The mini-bioreactor platform comprises a set of parallel shake flasks operating in fed-batch mode. The advantage of this system is its high experimental throughput for the evaluation of the GFP synthesis capacity of our reporter strains. In the case of classical shake flask system, no significant evolution of GFP synthesis have been observed, considering the reduced microbial growth period allowed by the system, whereas in the case of fed-batch operated mini-bioreactors, evolution of GFP synthesis, as well as GFP distribution among the microbial population, has been observed for three preselected strains (prpoS, puspA and posmC::gfp). More interestingly, a binary mode of expression has been observed in the case of the cultures carried out with the reporter strains for which GFP synthesis is under the control of the rpoS promoter which is induced under carbon limitation conditions. However, the generation of controlled glucose perturbations is relatively limited in this system and, in a second step fully automated bioreactor with a sclae-down strategy has been used to correlate the response of a prpoS::gfp strains with extracellular glucose perturbations. In the case of the culture performed in perturbed bioreactor (glucose intermittent feeding or glucose addition at the level of the recycle loop of a two-compartment scale-down bioreactor), the slowdown of the GFP synthesis resulting in the observation of a binary repartition of GFP content among the microbial population, has been observed. This observation led to the conclusion that the prpoS::gfp can be used as a biosensor for the validation of a fed-batch profile in industrial-scale bioreactors.     

Studying the non‐thermal effects of terahertz radiation on E. coli/pKatG‐GFP biosensor cells

Studies of the impact of terahertz radiation on living objects present a significant interest since its use for security systems is currently considered promising. We studied the non‐thermal impact of terahertz radiation on E. coli/pKatG‐gfp biosensor cells. The Novosibirsk free electron laser (NovoFEL), which currently has the world's highest average and peak power, was used as the source of terahertz radiation. We demonstrated that exposure to terahertz radiation at the wavelengths of 130, 150, and 200 µm and a power of 1.4 W/cm² induces changes in green fluorescent protein (GFP) fluorescence values and thus induces the expression of GFP in E. coli/pKatG‐gfp biosensor cells. Possible mechanisms of the E. coli response to non‐thermal exposure to terahertz radiation are discussed. Bioelectromagnetics 34:15–21, 2013. © 2012 Wiley Periodicals, Inc.     

Sporisorium scitamineum colonisation of sugarcane genotypes susceptible and resistant to smut revealed by GFP‐tagged strains

Sporisorium scitamineum is the causal agent of sugarcane smut disease. The fungus establishes a biotrophic interaction with sugarcane tissues, and unlike smut fungi of other monocot hosts, the primary meristem of sugarcane plants develops a whip‐like structure instead of a tumour‐like galls emerging from floral structures (tassels and ears). We examined (GFP)‐tagged S. scitamineum infecting tissues of three sugarcane genotypes with distinct responses to smut (susceptible, intermediate resistant and resistant). Mating compatible haploid cells gfp‐expressing were obtained by Agrobacterium tumefaciens‐mediated transformation (ATMT) using the integrative vector pFAT‐gfp. Regardless of the inoculation method (drop inoculation and hypodermal syringe inoculation), all genotypes were colonised by the fungus. GFP‐tagged strains of opposite mating reaction were able to: (a) grow in vitro as fluorescent yeast‐like cells; (b) generate infectious dikaryon; (c) penetrate sugarcane tissues; (d) colonise tissues by growing a filamentous network; and (e) form the characteristic highly branched hyphae within host cells. Fungal colonisation 160 DAI revealed an association of the fungus with vascular vessels disrupting their organisation in all three genotypes analysed. However, the resistant plants did not develop whips spanning the experiment time. The first whips emerged 76 DAI from plants of the susceptible genotype whereas for intermediate resistant plants whips were detected at 137 DAI. These whips were dissected and fluorescent sporogenesis and teliospore maturation were analysed. In vitro germination of recovered teliospores revealed after meiosis the formation of a three‐celled hyphal filament, where the fourth cell was likely maintained in the teliospore coat. These cells showed independent segregation of the gfp marker, as a result of gfp insertions in different chromosomes of each compatible haploid strain. This work presents the complete fungal life cycle of GFP‐marked S. scitamineum to study developmental stages in planta.     

Scale-down prediction of industrial scale pleated membrane cartridge performance

Flat‐sheet membrane discs represent the current standard format used for experimental prediction of the scale‐up of normal flow filtration processes. Use of this format is problematic, however, since the scale‐down results typically show a 40–55% difference in performance compared to large‐scale cartridges depending upon the feedstock used. In this work, novel pleated scale‐down devices (A_m = 1.51–15.1 × 10⁻³ m²) have been designed and fabricated. It is shown that these can more accurately predict the performance of industrial scale single‐use pleated membrane cartridges (A_m = 1.06 m²) commonly used within biopharmaceutical manufacture. The single‐use scale‐down cartridges retain the same pleat characteristics of the larger cartridges, but require a reduced feed volume by virtue of a substantially diminished number of active membrane pleats. In this study, a 1,000‐fold reduction in feed volume requirement for the scale‐down cartridge with the smallest membrane area was achieved. The scale‐down cartridges were tested both with clean water and a pepsin protein solution, showing flux‐time relationships within 10% of the large‐scale cartridge in both cases. Protein transmission levels were also in close agreement between the different scale cartridges. The similarity in performance of the scale‐down and the large‐scale cartridges, coupled with the low feed requirement, make such devices an excellent method by which rapid scale‐up can be achieved during early stage process development for biopharmaceutical products. This new approach is a significant improvement over using flat‐sheet discs as the quantitative similarity in performance with the large‐scale leads to reliable scale‐up predictions while requiring especially small volumes of feed material. Biotechnol. Bioeng. 2011; 108:830–838. © 2010 Wiley Periodicals, Inc.     

Heterologous expression of green fluorescent protein in Paenibacillus larvae, the causative agent of American Foulbrood of honey bees

Aims: We aimed at expressing heterologous proteins in Paenibacillus larvae, the causative agent of American Foulbrood of honey bees, as a prerequisite for future studies on the molecular pathogenesis of P. larvae infections. Methods and Results: For this purpose, we established a protocol for the transformation of the plasmid pAD43‐25 carrying a functional GFP gene sequence (gfpmut3a) into different P. larvae strains representing the two most relevant P. larvae genotypes ERIC I and ERIC II. We determined the optimal field strength for electroporation and the optimal regeneration time after transformation. Stable GFP expression could be detected in the mutants during their entire life cycles and even after sporulation and re‐germination. Conclusions: This method is suitable not only for the expression of GFP in P. larvae but also for the expression of heterologous proteins or GFP‐tagged proteins in P. larvae. Mutants can be used for infection assays because GFP expression remained stable after sporulation and re‐germination. Significance and Impact of the Study: This method provides the first true molecular tool for P. larvae and, therefore, is an immense advancement from what we had previously at our hands for the study of P. larvae pathogenesis.     

Lipid synthesized by micro‐algae grown in laboratory‐ and industrial‐scale bioreactors

Tetraselmis sp. and Nannochloropsis oculata, cultivated in industrial‐scale bioreactors, produced 2.33 and 2.44% w/w lipid (calculated as the sum of fatty acid methyl esters) in dry biomass, respectively. These lipids contained higher amounts of neutral lipids and glycolipids plus sphingolipids, than phospholipids. Lipids of Tetraselmis sp. were characterized by the presence of eicosapentaenoic acid (that was located mainly in phospholipids), and octadecatetraenoic acid (that was equally distributed among lipid fractions), while these fatty acids were completely absent in N. oculata lipids. Additionally, lipids produced by 16 newly isolated strains from Greek aquatic environments (cultivated in flask reactors) were studied. The highest percentage of lipids was found in Prorocentrum triestinum (3.69% w/w) while the lowest in Prymnesium parvum (0.47% w/w). Several strains produced lipids rich in eicosapentaenoic and docosahexaenoic acids. For instance, docosahexaenoic acid was found in high percentages in lipids of Amphidinium sp. S1, P. parvum, Prorocentrum minimum and P. triestinum, while lipids produced by Asterionella sp. (?) S2 contained eicosapentaenoic acid in high concentration. These lipids, containing ω‐3‐long‐chain polyunsaturated fatty acids, have important applications in the food and pharmaceutical industries and in aquaculture.     

Analyzing notochord segmentation and intervertebral disc formation using the twhh:gfp transgenic zebrafish model

To characterize the process of vertebral segmentation and disc formation in living animals, we analyzed tiggy-winkle hedgehog (twhh):green fluorescent protein (gfp) and sonic hedgehog (shh):gfp transgenic zebrafish models that display notochord-specific GFP expression. We found that they showed distinct patterns of expression in the intervertebral discs of late stage fish larvae and adult zebrafish. A segmented pattern of GFP expression was detected in the intervertebral disc of twhh:gfp transgenic fish. In contrast, little GFP expression was found in the intervertebral disc of shh:gfp transgenic fish. Treating twhh:gfp transgenic zebrafish larvae with exogenous retinoic acid (RA), a teratogenic factor on normal development, resulted in disruption of notochord segmentation and formation of oversized vertebrae. Histological analysis revealed that the oversized vertebrae are likely due to vertebral fusion. These studies demonstrate that the twhh:gfp transgenic zebrafish is a useful model for studying vertebral segmentation and disc formation, and moreover, that RA signaling may play a role in this process.     

Faithful expression of green fluorescent protein (GFP) in transgenic zebrafish embryos under control of zebrafish gene promoters

Although the zebrafish has become a popular model organism for vertebrate developmental and genetic analyses, its use in transgenic studies still suffers from the scarcity of homologous gene promoters. In the present study, three different zebrafish cDNA clones were isolated and sequenced completely, and their expression patterns were characterized by whole‐mount in situ hybridization as well as by Northern blot hybridization. The first clone encodes a type II cytokeratin (CK), which is specifically expressed in skin epithelia in early embryos and prominently expressed in the adult skin tissue. The second clone is muscle specific and encodes a muscle creatine kinase (MCK). The third clone, expressed ubiquitously in all tissues, is derived from an acidic ribosomal phosphoprotein P0 (arp) gene. In order to test the fidelity of zebrafish embryos in transgenic expression, the promoters of the three genes were isolated using a rapid linker‐mediated PCR approach and subsequently ligated to a modified green fluorescent protein (gfp) reporter gene. When the three hybrid GFP constructs were introduced into zebrafish embryos by microinjection, the three promoters were activated faithfully in developing zebrafish embryos. The 2.2‐kb ck promoter was sufficient to direct GFP expression in skin epithelia, although a weak expression in muscle was also observed in a few embryos. This pattern of transgenic expression is consistent with the expression pattern of the endogenous cytokeratin gene. The 1.5‐kb mck promoter/gfp was expressed exclusively in skeletal muscles and not elsewhere. By contrast, the 0.8‐kb ubiquitous promoter plus the first intron of the arp gene were capable of expressing GFP in a variety of tissues, including the skin, muscle, lens, neurons, notochord, and circulating blood cells. Our experiments, therefore, further demonstrated that zebrafish embryos can faithfully express exogenously introduced genes under the control of zebrafish promoters. Dev. Genet. 25:158–167, 1999. © 1999 Wiley‐Liss, Inc.     

Feeding strategies enhance high cell density cultivation and protein expression in milliliter scale bioreactors

Miniature bioreactors under parallel fed‐batch operations are not only useful screening tools for bioprocess development but also provide a suitable basis for eventual scale‐up. In this study, three feeding strategies were investigated: besides the established intermittent feeding by a liquid handler, an optimized microfluidic device and a new enzymatic release system were applied for parallel fed‐batch cultivation of Escherichia coli HMS174(DE3) and BL21(DE3) strains in stirred‐tank bioreactors on a 10 mL scale. Lower fluctuation in dissolved oxygen (DO) and higher optical densities were measured in fed‐batch processes applying the microfluidic device or the enzymatic glucose/fructose release system (conversion of intermittently added sucrose by an invertase), but no difference in dry cell weights (DCW) were observed. With all three feeding strategies high cell densities were realized on a milliliter scale with final optical density measured at 600 nm (OD₆₀₀) of 114–133 and final DCW concentrations of 69–70 g L^–1. The effect of feeding strategies on the expression of two heterologous proteins was investigated. Whereas no impact was observed on the expression of the spider silk protein eADF4(C16), the fluorescence of enhanced green fluorescence protein (eGFP) was reproducibly lower, if an intermittent glucose feed was applied. Thus, the impact of feeding strategy on expression is strongly dependent on the E. coli strain and/or expressed protein. As a completely continuous feed supply is difficult to realize in miniature bioreactors, the enzymatic release approach from this study can be easily applied in all microfluidic system to reduce fluctuations of glucose supply and DO concentrations.