Derivation and characterization of cholesterol-independent non-GS NS0 cell lines for production of recombinant antibodies

Presented is an antibody production platform based on the fed-batch culture of recombinant NS0-derived cell lines. NS0 host cells, obtained from the European Collection of Cell Cultures (ECACC, Salisbury, UK, Part No. 85110503), were first adapted to grow in a protein-free, cholesterol-free medium. The resulting host cell line was designated NS0-PFCF (protein-free, cholesterol-free). The five production cell lines presented here were generated using a common protocol consisting of transfection by electroporation and subcloning. The NS0-PFCF host cell line was transfected using a single expression vector containing the Escherichia coli xanthine-guanine phosphoribosyl transferase gene (gpt), and the antibody heavy and light chain genes driven by the CMV promoter. The five cell lines were chosen after one to three rounds of iterative subcloning, which resulted in a 19-64% increase in antibody productivity when four mother-daughter cell pairs were cultured in a fed-batch bioreactor process. The production cell lines were genetically characterized to determine antibody gene integrity, nucleotide sequences, copy number, and the number of insertion sites in the NS0 cell genome. Genetic characterization data indicate that each of the five production cell lines has a single stably integrated copy of the antibody expression vector, and that the antibody genes are correctly expressed. Stability of antibody production was evaluated for three of the five cell lines by comparing the early stage seed bank with the Working Cell Bank (WCB). Antibody productivity was shown to be stable in two of three cell lines evaluated, while one of the cell lines exhibited a 20% drop in productivity after passaging for approximately 4 weeks. These five NS0-derived production cell lines were successfully cultured to produce antibodies with acceptable product quality attributes in a standardized fed-batch bioreactor process, consistently achieving an average specific productivity of 20-60 pg/cell-day, and a volumetric productivity exceeding 120 mg/L-day (Burky et al., 2006). In contrast to the commonly available NS0 host cell line, which requires serum and cholesterol for growth, and the commonly used expression vector system, which uses a proprietary glutamine synthetase selection marker (GS-NS0), these NS0 cells are cholesterol-independent, grow well in a protein-free medium, use a non-proprietary selection marker, and do not require gene amplification for productivity improvement. These characteristics are advantageous for use of this NS0 cell line platform for manufacturing therapeutic antibodies.     

Protein-free fed-batch culture of non-GS NS0 cell lines for production of recombinant antibodies

Presented is a novel antibody production platform based on the fed-batch culture of recombinant, NS0-derived cell lines. A standardized fed-batch cell culture process was developed for five non-GS NS0 cell lines using enriched and optimized protein-free, cholesterol-free, and chemically defined basal and feed media. The process performed reproducibly and scaled faithfully from the 2-L to the 100-L bioreactor scale achieving a volumetric productivity of > 120 mg/L per day. Fed-batch cultures for all five cell lines exhibited significant lactate consumption when the cells entered the stationary or death phase. Peak and final lactate concentrations were low relative to a previously developed fed-batch process (FBP). Such low lactate production and high lactate consumption rates were unanticipated considering the fed-batch culture basal medium has an unconventionally high initial glucose concentration of 15 g/L, and an overall glucose consumption in excess of 17 g/L. The potential of this process platform was further demonstrated through additional media optimization, which has resulted in a final antibody concentration of 2.64 +/- 0.19 g/L and volumetric productivity of > 200 mg/L per day in a 13-day FBP for one of the five production cell lines. Use of this standardized protein-free, cholesterol-free NS0 FBP platform enables consistency in development time and cost effectiveness for manufacturing of therapeutic antibodies.     

Defined protein and animal component-free NS0 fed-batch culture

A chemically defined protein and animal component-free fed-batch process for an NS0 cell line producing a human IgG(1) antibody has been developed. The fed-batch feed profile was optimised in a step-wise manner. Depletion of measurable compounds was determined by direct analysis. The cellular need for non-measurable compounds was tested by continued culturing of cell suspension, removed from the bioreactor, in shake-flasks supplemented with critical substances. In the final fed-batch culture, 8.4 x 10(6) viable cells mL(-1) and 625 mg antibody L(-1) was obtained as compared to 2.3 x 10(6) cells mL(-1) and 70 mg antibody L(-1) in batch. The increase in cell density, in combination with a prolonged declining phase where antibody formation continued, resulted in a 6.2-fold increase in total cell yield, a 10.5-fold increase in viable cell hours and an 11.4-fold increase in product yield. These improvements were obtained by using a feed with glucose, glutamine, amino acids, lipids, sodium selenite, ethanolamine and vitamins. Specifically, supplementation with lipids (cholesterol) had a drastic effect on the maximum viable cell density. Calcium, magnesium and potassium were not depleted and a feed also containing iron, lithium, manganese, phosphorous and zinc did not significantly enhance the cell yield. The growth and death profiles in the final fed-batch indicated that nutrient deprivation was not the main cause of cell death. The ammonium concentration and the osmolality increased to potentially inhibitory levels, but an imbalance in the supply of growth/survival factors may also contribute to termination of the culture.     

A fully defined,fed‐batch,recombinant NS0 culture process for monoclonal antibody production

To manufacture a glycoprotein, mammalian cells expressing the desired protein are often grown in fed‐batch mode. Feeding an undefined, nonanimal hydrolysate helps the cells receive sufficient nutrition, but makes systems difficult to optimize. Even different lots of the same hydrolysate may have significant variability; furthermore, individual components may actually be detrimental to the cells. Switching to fully defined feeds could eliminate these issues. For monoclonal antibody (mAb) production by fed‐batch NS0 cells, this article describes the replacement of a hydrolysate‐based feed with a fully defined, animal‐component‐free feed system. The defined feed initially had 67 components, but additional experiments allowed a reduction to 25 components. The mAb titer is approximately 20% higher than in the undefined system, and the feed volume is circa 20% lower. The two systems generated antibodies with similar glycosylation profiles. Other benefits of the defined feed system include lower raw material costs, the ability to optimize key nutrient concentrations, greater confidence in raw material quality, and the elimination of potential, hydrolysate‐associated endotoxin issues. © 2010 American Institute of Chemical Engineers Biotechnol. Prog., 2010     

NS0 cell damage by high gas velocity sparging in protein-free and cholesterol-free cultures

Recent developments in high cell density and high productivity fed-batch animal cell cultures have placed a high demand on oxygenation and carbon dioxide removal in bioreactors. The high oxygen demand is often met by increasing agitation and sparging rates of air/O2 in the bioreactors. However, as we demonstrate in this study, an increase of gas sparging can result in cell damage at the sparger site due to high gas entrance velocities. Previous studies have showed that gas bubble breakup at the culture surface was primarily responsible for cell damage in sparged bioreactors. Such cell damage can be reduced by use of surfactants such as Pluronic F-68 in the culture. In our results, where NS0 cells were grown in a protein-free and cholesterol-free medium containing 0.5 g/L Pluronic F-68, high gas entrance velocity at the sparger site was observed as the second mechanism for cell damage. Experiments were performed in scaled-down spinners to model the effect of hydrodynamic force resulting from high gas velocities on antibody-producing NS0 cells. Cell growth and cell death were described by first-order kinetics. Cell death rate constant increased significantly from 0.04 to 0.18 day(-1) with increasing gas entrance velocity from 2.3 to 82.9 m/s at the sparger site. The critical gas entrance velocity for the NS0 cell line studied was found to be approximately 30 m/s; velocities greater than 30 m/s caused cell damage which resulted in reduced viability and consequently reduced antibody production. Observations from a second cholesterol-independent NS0 cell line confirmed the occurrence of cell damage due to high gas velocities. Increasing the concentration of Pluronic F-68 from 0.5 to 2 g/L had no additional protective effect on cell damage associated with high gas velocity at the sparger. The results of gas velocity analysis for cell damage have been applied in two case studies of large-scale antibody manufacturing. The first is a troubleshooting study for antibody production carried out in a 600 L bioreactor, and the second is the development of a gas sparger design for a large bioreactor scale (e.g., 10,000 L) for antibody manufacturing.     

Dynamic analysis of GS-NS0 cells producing a recombinant monoclonal antibody during fed-batch culture

In this study we have analyzed the dynamic covariation of the mammalian cell proteome with respect to functional phenotype during fed-batch culture of NS0 murine myeloma cells producing a recombinant IgG(4) monoclonal antibody. GS-NS0 cells were cultured in duplicate 10 L bioreactors (36.5 degrees C, 15% DOT, pH 7.0) for 335 h and supplemented with a continuous feed stream after 120 h. Cell-specific growth rate declined continuously after 72 h of culture. Cell-specific recombinant monoclonal antibody production rate (qP) varied sixfold through culture. Whilst qP correlated with relative recombinant heavy chain mRNA abundance up to 216 h, qP subsequently declined, independent of recombinant heavy chain or light chain mRNA abundance. GS-NS0 cultures were sampled at 48 h intervals between 24 and 264 h of culture for proteomic analyses. Total protein abundance and nascent polypeptide synthesis was determined by 2D PAGE of unlabeled proteins visualized by SYPRO Ruby and autoradiography of (35)S-labeled polypeptides, respectively. Covariation of nascent polypeptide synthesis and abundance with biomass-specific cell growth, glucose and glutamate consumption, lactate and Mab production rates were then examined using two partial least squares regression models. Most changes in polypeptide synthesis or abundance for proteins previously identified by mass spectrometry were positively correlated with biomass-specific growth rate. We conclude that the substantial transitions in cell physiology and qP that occur during culture utilize a relatively constant complement of the most abundant host cell machines that vary primarily with respect to induced changes in cell growth rate.     

Molecular analysis of successful cell line selection in transfected GS-NS0 myeloma cells

The production of recombinant proteins from mammalian cells is now an essential part of biotechnology. However, despite this importance, the detailed characteristics of good producing cell lines remain largely unknown. The industrially important GS-NS0 mammalian expression system is able to produce large amounts of protein from relatively few copies of recombinant genes. This makes GS-NS0 cell lines ideal candidates to study the consequence of recombinant plasmid transfection in mammalian cells. This study investigated the molecular features of a panel of 17 randomly chosen GS-NS0 cell lines engineered to produce a recombinant antibody. The research analysed antibody production via enzyme-linked immunosorbent assay (ELISA), and investigated the molecular features of the transfectants by Northern, Southern and copy number analysis. The cell lines generated produced a range of antibody concentrations. In addition, for transfectants defined as producers of recombinant antibody there was a positive correlation between specific productivity and heavy chain mRNA expression. The use of Northern and Southern analysis allowed determination of the functional integrity of the transfected plasmid. Over 50% of the transfectants studied had molecular defects at the level of mRNA and/or cDNA. Cell lines were identified with suspected defects in the regulatory regions of transfected genes in addition to cell lines which lacked recombinant genes. Also, "false-positive" cell lines were generated which were able to overcome the GS selection pressure without producing any recombinant antibody. This article discusses these findings in relation to vector design.     

Fed-batch bioreactor process scale-up from 3-L to 2,500-L scale for monoclonal antibody production from cell culture

This case study focuses on the scale-up of a Sp2/0 mouse myeloma cell line based fed-batch bioreactor process, from the initial 3-L bench scale to the 2,500-L scale. A stepwise scale-up strategy that involved several intermediate steps in increasing the bioreactor volume was adopted to minimize the risks associated with scale-up processes. Careful selection of several available mixing models from literature, and appropriately applying the calculated results to our settings, resulted in successful scale-up of agitation speed for the large bioreactors. Consideration was also given to scale-up of the nutrient feeding, inoculation, and the set-points of operational parameters such as temperature, pH, dissolved oxygen, dissolved carbon dioxide, and aeration in an integrated manner. It has been demonstrated through the qualitative and the quantitative side-by-side comparison of bioreactor performance as well as through a panel of biochemical characterization tests that the comparability of the process and the product was well controlled and maintained during the process scale-up. The 2,500-L process is currently in use for the routine clinical production of Epratuzumab in support of two global Phase III clinical trials in patients with lupus. Today, the 2,500 L, fed-batch production process for Epratuzumab has met all scheduled batch releases, and the quality of the antibody is consistent and reproducible, meeting all specifications, thus confirming the robustness of the process.     

Adaptation of cholesterol-requiring NS0 mouse myeloma cells to high density growth in a fully defined protein-free and cholesterol-free culture medium

NS0 has been used as a fusion partner for the production of hybridomas and has more recently been engineered to produce recombinant protein. A protein-free culture medium, designated W38 medium, has previously been developed which supported high density growth of rat myeloma and hybridoma cell lines. NS0 cells failed to grow in W38 medium and in a number of protein-free culture media which support the growth of other myeloma cell lines. NS0 cells are derived from the NS-1 cell line, which is known to require exogencus cholesterol. It was found that NS0 cells grew in W38 medium supplemented with phosphatidylcholine, cholesterol, and albumin and that NS0 were auxotrophic for cholesterol. Protein-free growth of NS0 cells was achieved by using -cyclodextrin to replace albumin as a lipid carrier. The maximal cell density reached in this protein-free medium was in excess of 1.5×10⁶ cell ml^–1. The lipid supplements in the medium precipitated after a few days storage at +4°C. In order to overcome this problem a protocol was developed which allowed NS0 cells to be adapted to cholesterol-independent growth in W38 medium. NS0.CF (cholesterol-independent NS0 cells) were cultured continuously in W38 medium for several months. In shake flask culture a cell density of 2.4×10⁶ cells ml^–1 was achieved in W38 medium compared with 1.41×10⁶ cells ml^–1 in RPMI 1640 medium containing 10% foetal bovine serum. NS0.CF cells readily grew in a 1 litre stirred bioreactor using W38 medium supplemented with Pluronic F68 reaching a density of 3.24×10⁶ cells ml^–1. NS0.CF were cloned protein-free by limiting dilution in W38 medium, giving colonies in wells that were seeded at an average density of 0.32 cells per 200 l. This study has demonstrated for the first time the growth of a cholesterol-requiring mouse myeloma cell line in a completely defined protein-free medium and its subsequent adaptation to cholesterol-independence.Abbreviations BSA bovine serum albumin - C cholesterol - CD cyclodextrin - F68 Pluronic F68 - GS glutamine synthetase - P phosphatidylcholine - PC-FBS phosphatidylcholine, cholesterol and foetal bovine serum - RPMI RPMI 1640 medium - MSX methionine sulphoximine     

The response of GS-NS0 myeloma cells to single and multiple pH perturbations

Animal cells are cultured in several types of vessels at laboratory and industrial scale the most common being the stirred tank and the air-lift. Economically, it is preferable to culture animal cells at the largest possible scale but the perceived sensitivity of animal cells to hydrodynamic shear has, until now, limited the aeration and agitation rates used. This has been reported to cause inhomogeneities in operational parameters such as dissolved oxygen concentration, temperature and pH. pH is of special interest during the latter stages of many animal cell fermentation because alkali additions, used for pH control, can cause large local pH perturbations of varying size and duration. The effect of single and multiple pH perturbations on the cell growth of a widely used GS-NS0 mouse myeloma cell line grown in batch culture was investigated. The effect of perturbation amplitude and duration was investigated using a single stirred tank reactor (STR). In the single STR system cells were subjected to one pH 8.0 or 9.0 perturbation ranging in duration from 0-90 minutes. No measurable decrease in viable cell number was seen for pH 8.0 perturbations of any duration whereas pH 9.0 perturbations lasting for 10 minutes caused a 15% decrease in viable cell number. The proportion of viable cells decreased with increasing perturbation time and a 90-minute exposure killed all of the cells. The effect of multiple pH perturbations on GS-NS0 cells was investigated using two connected STR's. More specifically the number of perturbations and the perturbation frequency were investigated. Cells were subjected to between 0 and 100 perturbations at pH 8.0; the time between each perturbation (frequency) was 6 minutes and each perturbation lasted for 200 seconds. Viable cell number decreased with increasing perturbation number, with 100 perturbations causing death of 27.5% of cells. Cells were also exposed to 10 perturbations at pH 9.0, each of 200 second duration at frequencies of either 6, 18 or 60 minutes. Approximately 8 times more cells were killed with perturbations at a 6-minute frequency (28.3% cell death) than at a 60-minute frequency (3.4% cell death).     

Suppression of apoptosis in perfusion culture of Myeloma NS0 cells enhances cell growth but reduces antibody productivity

A spin filter perfusion systems was used to achieve a high cell density culture for two NS0 cell lines in 2 litres bioreactors. One cell line is transfected with the bcl-2 gene (NS0 Bcl-2) encodes the 'anti-apoptotic' human Bcl-2 protein and the other cell line (NS0 Control) with a blank vector. The runs started as batch cultures for two days and were perfused with fresh medium at 0.5 volumes per day (day(-1)) for 4 days, increasing gradually to 2 day(-1) at day 7. The increase of the viable cell density of Bcl-2 cell line was far greater than the control cell line, although they were perfused with the same amount of medium. At the end of the period of each perfusion rate, the viable cell densities of Bcl-2 culture were 30%, 120%, 160% and 220% higher than its control cell line corresponding values. Overall, there was a roughly 9 fold increase in viable cell density from the inoculum for the control culture, but almost a 30 fold increase for the Bcl-2 culture. The mode of cell death in the control culture was initially predominantly by necrosis (viability higher than 80%), but apoptotic cell death became more significant after day 8 of the culture. Cell death in the Bcl-2 culture was almost entirely by necrosis, although it remained at a very low level (less than 5%) to the termination time. The cell cycle distributions for both cell lines were very much similar indicating they have a similar doubling time and G1 to S progression rate. Interestingly, the Bcl-2 cultures exhibited reduced antibody specific production rate with increasing viable cell number and time. The volumetric production rate was, however, similar in both cultures. Bcl-2 as an anti-death protein allowed cells to survive and thus divide to higher cell densities without the need for additional nutrients. Most of the cellular energy in a producer cell line is used for biomass production rather than for antibody production, as was the case with the control cell line.     

The response of GS-NS0 myeloma cells to pH shifts and pH perturbations

The perceived sensitivity of animal cells to hydrodynamic shear has limited agitation and aeration at large-scale. This makes it difficult to ensure adequate mixing of the vessel contents and may lead to inhomogeneities in operational parameters such as temperature, dissolved oxygen concentration, and especially pH. The effect of pH shifts and pH perturbations on the cellular responses, in batch culture, of a GS-NS0 mouse myeloma cell line, expressing a recombinant antibody, was investigated. In addition, the effect of extreme pH on the structure of the purified antibody product was studied using isoelectric focusing. The fermentation pH value was shifted abruptly from pH 7.3 to pH values ranging from 6.5 to 9.0. Culture pH was maintained at this new value for the remainder of the fermentation. All pH shifts of above 0.2 units caused a transient increase in apoptosis. However, cultures shifted to pH values between 7.0 and 8.0 continued to grow and the apoptotic fraction returned to initial levels. Cultures shifted to pH values above pH 8.0 and below pH 7.0 did not recover resulting in culture death. For example, a shift to pH 8.5 caused accumulation of cells in the G(2)/M phase of the cell cycle followed by apoptotic death. After the pH shift, maximum specific growth rate was observed over the range pH 7.3 to 7.5 and maximum viable cell number was seen at pH 7.3. Maximum volumetric antibody production, resulting from increased culture longevity, was seen at pH 7.0. It was also observed that glucose consumption increased with increasing pH. In a separate set of experiments cells were subjected to a single pH perturbation ranging in duration from 0 to 600 minutes. Exposure of cells to a pH value greater than 8.5 for more than 10 minutes caused a decrease in the proportion of viable cells and induced a lag in cell growth. At very low pH (6.5) similar effects were seen, but only for extended perturbations (600 min). However, after recovery from the pH perturbation, growth, product secretion and metabolism all returned to original levels. Incubation of the antibody, at the range of pH values investigated, indicated no alterations in the structure of the antibody as determined by the isoelectric focusing pattern.     

Integrated continuous bioprocessing: Economic,operational, and environmental feasibility for clinical and commercial antibody manufacture

This paper presents a systems approach to evaluating the potential of integrated continuous bioprocessing for monoclonal antibody (mAb) manufacture across a product's lifecycle from preclinical to commercial manufacture. The economic, operational, and environmental feasibility of alternative continuous manufacturing strategies were evaluated holistically using a prototype UCL decisional tool that integrated process economics, discrete‐event simulation, environmental impact analysis, operational risk analysis, and multiattribute decision‐making. The case study focused on comparing whole bioprocesses that used either batch, continuous or a hybrid combination of batch and continuous technologies for cell culture, capture chromatography, and polishing chromatography steps. The cost of goods per gram (COG/g), E‐factor, and operational risk scores of each strategy were established across a matrix of scenarios with differing combinations of clinical development phase and company portfolio size. The tool outputs predict that the optimal strategy for early phase production and small/medium‐sized companies is the integrated continuous strategy (alternating tangential flow filtration (ATF) perfusion, continuous capture, continuous polishing). However, the top ranking strategy changes for commercial production and companies with large portfolios to the hybrid strategy with fed‐batch culture, continuous capture and batch polishing from a COG/g perspective. The multiattribute decision‐making analysis highlighted that if the operational feasibility was considered more important than the economic benefits, the hybrid strategy would be preferred for all company scales. Further considerations outside the scope of this work include the process development costs required to adopt continuous processing. © 2017 The Authors Biotechnology Progress published by Wiley Periodicals, Inc. on behalf of American Institute of Chemical Engineers Biotechnol. Prog., 33:854–866, 2017     

Generation of high rapamycin producing strain via rational metabolic pathway‐based mutagenesis and further titer improvement with fed‐batch bioprocess optimization

Rapamycin is a triene macrolide antibiotic produced by Streptomyces hygroscopicus. Besides its wide application as an effective immunosuppressive agent, other important bioactivities have made rapamycin a potential drug lead for novel pharmaceutical development. However, the low titer of rapamycin in the original producer strain limits further industrialization efforts and restricts its use for other applications. Predicated on knowledge of the metabolic pathways related to rapamycin biosynthesis in S. hygroscopicus, we have rationally designed approaches to generate a rapamycin high producer strain of S. hygroscopicus HD‐04‐S. These have included alleviation of glucose repression, improved tolerance towards lysine and shikimic acid, and auxotrophy of tryptophan and phenylalanine through the application of stepwise UV mutagenesis. The resultant strain produced rapamycin at 450 mg/L in the shake flask scale. These fermentations were further scaled up in 120 and 20,000 L fermentors, respectively, at the pilot plant. Selected fermentation factors including agitation speed, pH, and on‐line supplementation were systematically evaluated. A fed‐batch strategy was established to maximize rapamycin production. With these efforts, an optimized fermentation process in the larger scale fermentor was developed. The final titer of rapamycin was 812 mg/L in the 120 L fermentor and 783 mg/L in the 20,000 L fermentor. This work highlights a high rapamycin producing strain derived by mutagenesis and subsequent screening, fermentation optimization of which has now made it feasible to produce rapamycin on an industrial scale by fermentation. The strategies developed here should also be applicable to titer improvement of other important microbial natural products on an industrial scale. Biotechnol. Bioeng. 2010;107: 506–515. © 2010 Wiley Periodicals, Inc.     

Use of a Plackett-Burman statistical design to determine the effect of selected amino acids on monoclonal antibody production in CHO cells

Culture media design is central to the optimization of monoclonal antibody (mAb) production. Although general strategies do not currently exist for optimization of culture media, the combined use of statistical design and analysis of experiments and strategies based on simple material balances can facilitate culture media design. In this study, we evaluate the effect of selected amino acids on the growth rate and monoclonal antibody production of a Chinese hamster ovary DG-44 (CHO-DG44) cell line. These amino acids were selected based on their relative mass fraction in the specific mAb produced in this study, their consumption rate during bioreactor experiments, and also through a literature review. A Plackett-Burman statistical design was conducted to minimize the number of experiments needed to obtain statistically relevant information. The effect of this set of amino acids was evaluated during exponential cell culture (considering viable cell concentration and the specific growth rate as main output variables) and during the high cell-density stage (considering mAb final concentration and specific productivity as relevant output variables). For this particular cell line, leucine (Leu) and arginine (Arg) had the highest negative and positive effects on cell viability, respectively; Leu and threonine (Thr) had the highest negative effect on growth rate, and valine (Val) and Arg demonstrated the highest positive impact on mAb final concentration. Results suggest the pertinence of a two-stage strategy for amino acid supplementation, with a mixture optimized for cell growth and a different amino acid mixture for mAb production at high density.     

Large-scale production of endotoxin-free plasmids for transient expression in mammalian cell culture

Transient expression of recombinant proteins in mammalian cell culture in a 100-L scale requires a large quantity of plasmid that is very labour intensive to achieve with shake flask cultures and commercially available plasmid purification kits. In this paper we describe a process for plasmid production in 100-mg scale. The fermentation is carried out in a 4-L fed-batch culture with a minimal medium. The detection of the end of batch and triggering the exponential (0.1 h(-1)) feed profile was unattended and controlled by Multi-fermenter Control System. A restricted specific growth rate in fed-batch culture increased the specific plasmid yield compared to batch cultures with minimal and rich media. This together with high biomass concentration (68-107 g L(-1) wet weight) achieves high volumetric yields of plasmid (95-277 mg L(-1) depending on the construct). The purification process consisted of alkaline lysis, lysate clarification and ultrafiltration, two-phase extraction with Triton X-114 for endotoxin removal, anion-exchange chromatography as a polishing step, ultrafiltration and sterile filtration. Both fermentation and purification processes were used without optimisation for production of four plasmids yielding from 39 to 163 mg of plasmids with endotoxin content of 2.5 EU mg(-1) or less.