Nutrient Feed Rate Research Articles

Optimization-based control strategy for a large-scale polyhydroxyalkanoates production in a fed-batch bioreactor using a coupled PDE–ODE system

Abstract Control strategy development for fed-batch bioreactor (FBBR) plays an important role in the improvement of polyhydroxyalkanoate (PHA) production. To develop a feeding strategy for PHA production in a large-scale FBBR, an optimization-based control scheme that considers nutrient dispersion is proposed in this work. A coupled partial differential equations and ordinary differential equation model is proposed to describe the axial-dispersed nutrient and well-dispersed microbial dynamics with process constraints. An analytical model predictive control (AMPC) method that applies integrated variables of nutrients is employed to develop the real-time control system. The control objective is to regulate the PHA concentration at the updated set points by adjusting the nutrient feed rates; a process disturbance is introduced to evaluate the control robustness. Simulation experiments of a fed-batch operation are conducted to investigate the performance of the developed controller; the controlled output is designed to track the updated set points corresponding to the biomass concentration. Results of closed-loop and regulatory systems showed that the proposed control strategy could provide more productivity (33–38%) compared to the applied PI controller. The performance test demonstrates that the developed control system could apply the biomass concentration for updating set points, provide the optimal control actions that promote PHB accumulation and handle the disturbance effectively.

Model‐based control of titer and glycosylation in fed‐batch mAb production: Modeling and control system development

AbstractTherapeutic monoclonal antibodies (mAbs) are typically manufactured using mammalian cell cultures in fed‐batch bioreactors, with increasing emphasis on meeting productivity and product quality attribute targets that depend strongly on such process variables as nutrient feed rates and bioreactor operating conditions. In this article, we identify, categorize, and address the challenges of achieving both productivity and product quality goals simultaneously, by developing a multivariable, model‐based control system that can satisfy multiple production objectives in a fed‐batch cell culture process. Here, we discuss model development and present theoretical concepts of observability and controllability that are essential to understanding and handling effectively these intrinsic challenges. Subsequently, we evaluate via simulation the performance of the outer‐loop model predictive control and demonstrate the overall capability to satisfy complex production objectives in a laboratory scale bioreactor, as a first step toward the ultimate goal of creating an advanced control system for fed‐batch mAb manufacturing processes.

At-line raman spectroscopy and design of experiments for robust monitoring and control of miniature bioreactor cultures.

The biopharmaceutical industry is moving toward a more quality by design (QbD) approach that seeks to increase product and process understanding and process control. Miniature bioreactor systems offer a high-throughput method enabling the assessment of numerous process variables in a controlled environment. However, the number of off/at-line samples that can be taken is restricted due to the small working volume of each vessel. This limitation may be resolved through the use of Raman spectroscopy due to its ability to obtain multianalyte data from small sample volumes fast. It can, however, be challenging to implement this technique for this application due to the complexity of the sample matrix and that analytes are often present in low concentration. Here, we present a design of experiments (DOE) approach to generate samples for calibrating robust multivariate predictive models measuring glucose, lactate, ammonium, viable cell concentration (VCC) and product concentration, for unclarified cell culture that improves the daily monitoring of each miniature bioreactor vessel. Furthermore, we demonstrate how the output of the glucose and VCC models can be used to control the glucose and main nutrient feed rate within miniature bioreactor cultures to within qualified critical limits set for larger scale vessels. The DOE approach used to generate the calibration sample set is shown to result in models more robust to process changes than by simply using samples taken from the "typical" process. © 2018 American Institute of Chemical Engineers Biotechnol. Prog., 35: e2740, 2019.

The use of dissolved oxygen-controlled, fed-batch aerobic cultivation for recombinant protein subunit vaccine manufacturing

A simple “off-the-shelf” fed-batch approach to aerobic bacterial cultivation for recombinant protein subunit vaccine manufacturing is presented. In this approach, changes in the dissolved oxygen levels are used to adjust the nutrient feed rate (DO-stat), so that the desired dissolved oxygen level is maintained throughout cultivation. This enables high Escherichia coli cell densities and recombinant protein titers. When coupled to a kLa-matched scale-down model, process performance is shown to be consistent at the 2L, 20L, and 200L scales for two recombinant E. coli strains expressing different protein subunit vaccine candidates. Additionally, by mining historical DO-stat nutrient feeding data, a method to transition from DO-stat to a pre-determined feeding profile suitable for larger manufacturing scales without using feedback control is demonstrated at the 2L, 20L, and 200L scales.

Carrying capacity for finfish aquaculture, Part II – Rapid assessment using hydrodynamic and semi-analytic solutions

The carrying capacity for aquaculture cage farming in Spencer Gulf (South Australia) is based in part on guidelines that the maximum feed rates and nutrient flux into the lease regions are determined such that the maximum nutrient concentration c does not exceed a prescribed value (say cP) to ensure ecosystem health – ecological carrying capacity. The goal of this study is to allow the rapid estimation of maximum nutrient fluxes and feed rates at new lease sites. Spencer Gulf is chosen as a case study although the methodology should find application in other regions around the world. In part I of this study, semi-analytic solutions were obtained to show that to a good approximation the maximum nutrient flux (feed rate) F can be simply estimated from: F=cP/T* where T* is a flushing time scale of the cage or lease region. In this study a 3-dimensional hydrodynamic model for Spencer Gulf is used to determine the parameters needed to estimate T* and thus F and feed rates at every model cell in the gulf. The parameters needed include the vector mean speed (U), r.m.s. tidal amplitude (UK) and the mean shear dispersion diffusivity (KS). As a case study, these parameters and T*, are estimated by three-monthly, winter averages. Results show the vector mean speed to be very small (U∼0.01m/s), tidal velocities large (UK∼0.3–1m/s) and the associated shear dispersion coefficients very large (KS∼10–100m2/s). Flushing at the scale of the lease (600m) and in the upper gulf is generally dominated by diffusive affects for which the maximum nutrient flux (and feed rates) is largest. The results should find application in other finite source flux problems in the coastal oceans including desalination plants and ocean outfalls.

Enhancing thermo-induced recombinant protein production in Escherichia coli by temperature oscillations and post-induction nutrient feeding strategies

Traditional strategies for production of thermo-induced recombinant protein in Escherichia coli consist of a two-phase culture, with an initial growth stage at low temperature (commonly 30°C) followed by a production stage where temperature is increased stepwise (commonly up to 42°C). A disadvantage of such strategies is that growth is inhibited upon temperature increase, limiting the duration of the production stage and consequently limiting recombinant protein production. In this work, a novel oscillatory thermo-induction strategy, consisting on temperature fluctuations between 37 and 42°C or 30 and 42°C, was tested for improving recombinant protein production. In addition, the induction schemes were combined with one of three different nutrient feeding strategies: two exponential and one linear. Recombinant human preproinsulin (HPPI), produced under control of the λPL-cI857 system in the E. coli BL21 strain, was used as the model protein. Compared to the conventional induction scheme at constant temperature (42°C), longer productive times were attained under oscillatory induction, which resulted in a 1.3- to 1.7-fold increase in maximum HPPI concentration. Temperature oscillations led to a 2.3- to 4.0-fold increase in biomass accumulation and a decrease of 48–62% in the concentration of organic acids, compared to conventional induction. Under constant induction, growth ceased upon temperature increase and the maximum concentration of HPPI was 3.9g/L, regardless of the post-induction feeding strategy used. In comparison, the combination of temperature oscillations and a high nutrient-feeding rate allowed sustained growth after induction and reaching up to 5.8g/L of HPPI.

Modelling and dynamic simulation of struvite precipitation from source-separated urine

A model of a mixed-mode nutrient recovery reactor is developed for a urine feed, incorporating complex solution thermodynamics, dynamic conservation relations and a power-law kinetic expression for crystal growth from seed crystals. Simulations at nominal operating conditions predict phosphorus recoveries greater than 99%, based on existing process kinetic parameters and operating conditions employed in previously published studies. The overall rate of nutrient recovery depends on the saturation index of the precipitating solid, the available surface area for mass transfer and the feed rate of the limiting constituent ion. Under the conditions considered, the nutrient feed rate appears to be the limiting factor for precipitation. Simulations demonstrate that diurnal feed flow variations of ±50% have a small effect on the rate of nutrient recovery. Overall, the study shows that valuable insights are gained in relation to process performance predictions, which should lead to more confident process design, operation and control.

Performance of a biological degradation method for indoor formaldehyde removal

Formaldehyde is a common gaseous pollutant affecting the indoor air quality and the health of occupants in buildings. Various technologies have been tried, such as adsorption, photocatalytic oxidization and botanical filtrations. However, the removal of indoor formaldehyde is still a challenging problem due to the low rate, byproduct formation and low efficiency of the above mentioned methods. Biological degradation technology with high removal efficiency has been successfully applied in industries for effluents and waste gas treatments. However, the potential biohazard risk from possible microorganism leakages of the biological system is the biggest concern of indoor air biofiltration systems. This study investigated the performance of a biological degradation system for formaldehyde removal at typical indoor condition. A microbial immobilization method was applied in the biofiltration system to reduce the potential biohazard risk. Bacterial strain of Pseudomonas putida was embedded in calcium alginate gel, which was then formed to a novel packing material called microbial granules. To investigate the operating performance of the biofilter in practical indoor environment, the concentrations of experimental inlet formaldehyde was controlled between 0.222 and 1.339 mg m−3, while the corresponding removal efficiency of the biofilter system varied from 68.6% to 93.5%. The pressure drop was maintained at 120 Pa with air flow rate increased up to 1.5 L min−1. The effect of inlet concentration, empty bed residence time, and nutrient feeding rate on the performance of the system was investigated as well. No microbial leakages were detected at air downstream during the experimental study, while the high microbial activity of bacteria entrapped in the packing material was observed.

Developing a scalable model of recombinant protein yield from Pichia pastoris: the influence of culture conditions, biomass and induction regime

BackgroundThe optimisation and scale-up of process conditions leading to high yields of recombinant proteins is an enduring bottleneck in the post-genomic sciences. Typical experiments rely on varying selected parameters through repeated rounds of trial-and-error optimisation. To rationalise this, several groups have recently adopted the 'design of experiments' (DoE) approach frequently used in industry. Studies have focused on parameters such as medium composition, nutrient feed rates and induction of expression in shake flasks or bioreactors, as well as oxygen transfer rates in micro-well plates. In this study we wanted to generate a predictive model that described small-scale screens and to test its scalability to bioreactors.ResultsHere we demonstrate how the use of a DoE approach in a multi-well mini-bioreactor permitted the rapid establishment of high yielding production phase conditions that could be transferred to a 7 L bioreactor. Using green fluorescent protein secreted from Pichia pastoris, we derived a predictive model of protein yield as a function of the three most commonly-varied process parameters: temperature, pH and the percentage of dissolved oxygen in the culture medium. Importantly, when yield was normalised to culture volume and density, the model was scalable from mL to L working volumes. By increasing pre-induction biomass accumulation, model-predicted yields were further improved. Yield improvement was most significant, however, on varying the fed-batch induction regime to minimise methanol accumulation so that the productivity of the culture increased throughout the whole induction period. These findings suggest the importance of matching the rate of protein production with the host metabolism.ConclusionWe demonstrate how a rational, stepwise approach to recombinant protein production screens can reduce process development time.

N-Hexane Biodegradation in Trickle Bed Air Biofilters

While hydrophilic compounds are degraded easily in Trickling bed air biofilters (TBABs), hydrophobic compounds are retarded until biological cultures produce a sufficient RNA or enzyme/protein to utilize this compound. Hydrophobic compounds are not readily bio-available which makes them reluctant to biodegradation as mass transfer between the gas and liquid phases is a rate limiting step. To enhance the destruction of hydrophobic compounds in TBABs, the utilization of surfactant was introduced to increase the solubility which helps overcoming the rate limiting step. The surfactant was used as well to limit the growth of excess biomass ensuring smooth flow through the biofilter bed and preventing short circuits. Two different non-ionic non-toxic surfactants were used in this study: Triton X-100 and Tomadol® 25-7. Two lab-scale controlled TBABs were operated for investigating the performance difference for n-Hexane as an example of hydrophobic volatile organic compound (VOC) with and without the addition of surfactant. Operating conditions in both TBABs were as follows: nutrient feed rate (2L/day), air flowrate (1.4L/min), bed depth (60cm), empty bed retention time (120s), bed material (diatomaceous earth pellets) and room-temperature. The inlet concentration was changed from 50 to 100ppmv. Acclimation period, removal profile along biofilter depth, nitrogen consumption, and CO2 production were compared under continuous loading operation condition. The optimum concentration of surfactant in the nutrient feed was determined by a batch experiment. The effect of different surfactant concentrations on VOC water solubility with time was studied by considering different VOC concentration sets within the TBAB loading rate range.

Effects of post-induction feed strategies on secretory production of recombinant streptokinase in Escherichia coli

The effects of post-induction nutrient feed rates, on recombinant streptokinase production in fed-batch processes, were investigated using various feed profiles. Recombinant streptokinase was produced using a secretory expression system and was induced by a temperature up-shift, using a temperature-sensitive λ PL promoter. The specific growth rates decreased sharply upon induction of recombinant protein expression, thus necessitating a variable feed strategy in the post-induction phase. The various feed profiles employed in the post-induction phase included constant feed rates, linearly increasing feed rate and exponentially varying feed rates. Significant differences were obtained in the specific activity of streptokinase produced in these fed-batch processes. A maximum activity per unit biomass of 4.96 × 10 6 and 4.43 × 10 6 IU/g DCW was achieved for exponentially decreasing feed and linearly increasing feed, respectively. The decrease in specific growth rate during the post-induction phase was also less pronounced in these cases in comparison to other fed-batch experiments. It was observed that streptokinase productivity is governed by the nutrient feed rate per unit biomass at a critical juncture after induction. The highest activity per unit biomass was obtained when the nutrient feed rate per unit biomass was around 0.7–0.8 g glucose (g DCW) −1 h −1, between 2 and 4 h after induction.

An on-line adaptive control based on DO/pH measurements and ANN pattern recognition model for fed-batch cultivation

An on-line adaptive control strategy based on DO/pH measurements and artificial neural network pattern recognition (ANNPR) model for fed-batch cultivation processes was proposed. Various changing patterns of pH and dissolved oxygen concentration (DO) under pH-Stat, DO-Stat, and the conditions of substrate in excess were collected and used to train the ANNPR models. Based on the on-line measured pH and DO data, the recognition results on current physiological state was deduced by the ANNPR models, and then the on-line adaptive control of nutrient feeding rate was implemented. Compared with the traditional pH-Stat control, the proposed control strategy increased cell productivity about 150% without byproduct accumulation. The control strategy is potentially useful for high cell density cultivation of recombinant microorganisms to efficiently express value-added foreign proteins or enzyme with the most traditional pH and DO sensors.

Characterization of the metabolic burden on Escherichia coli DH1 cells imposed by the presence of a plasmid containing a gene therapy sequence

The presence of a plasmid, containing gene sequences for DNA immunotherapy that are not expressed in microbial culture, imposed a degradation in bioreactor performance in cultures of the host E. coli strain. Significant decreases in growth rate (24%) and biomass yield (7%) and a corresponding increase in overflow metabolism were observed in a strain containing a therapeutic sequence (a hepatitis B antigen under the control of a CMV promotor). The observed increase in overflow metabolism was incorporated into a Metabolic Flux Analysis (MFA) model (as acetate secretion). Metabolic flux analysis revealed an increase in TCA cycle flux, consistent with an increased respiration rate observed in plasmid-containing cells. These effects are thought to result from increased ATP synthesis requirements (24%) arising from the expression of the Kanr plasmid marker gene whose product accounted for 18% of the cell protein of the plasmid-containing strain. These factors will necessitate significantly higher aeration and agitation rates or lower nutrient feed rates in high-density cultures than would be expected for plasmid-free cultures.

Design and Installation of a Next Generation Pilot Scale Fermentation System

Four new fermenters were designed and constructed for use in secondary metabolite cultivations, bioconversions, and enzyme production. A new PC/PLC-based control system also was implemented using GE Fanuc PLCs, Genius I/O blocks, and Fix Dynamics SCADA software. These systems were incorporated into an industrial research fermentation pilot plant, designed and constructed in the early 1980s. Details of the design of these new fermenters and the new control system are described and compared with the existing installation for expected effectiveness. In addition, the reasoning behind selection of some of these features has been included. Key to the design was the goal of preserving similarity between the new and previously existing and successfully utilized fermenter hardware and software installations where feasible but implementing improvements where warranted and beneficial. Examples of enhancements include strategic use of Inconel as a material of construction to reduce corrosion, piping layout design for simplified hazardous energy isolation, on-line calculation and control of nutrient feed rates, and the use of field I/O modules located near the vessel to permit low-cost addition of new instrumentation.

Actinorhodin and undecylprodigiosin production in wild-type and relA mutant strains of Streptomyces coelicolor A3(2) grown in continuous culture

The effects of growth rate and nutrient feed rate on the production of actinorhodin (Act) and undecylprodigiosin (Red) were determined in Streptomyces coelicolor A3(2) and in a congenic relA null-mutant known to be deficient in ppGpp synthesis and antibiotic production under conditions of nitrogen limitation. In the relA + strain, Act production was inversely related to specific growth rate in continuous cultures limited by glucose, ammonium, or phosphate, while Red biosynthesis was optimal at 0.05 h −1 regardless of the specific nutrient limitation. Production of Act and Red in the relA mutant was lower than that of the parental strain, particularly under conditions of glucose- and ammonium-limitation, indicating an important and general role for ppGpp in determining the onset of the antibiotic biosynthesis under conditions of nutrient limitation. At constant growth rate, but with varying nutrient feed rates, the specific rate of Act production was adversely influenced by increasing levels of glucose, ammonium, and phosphate, with phosphate having the greatest inhibitory effect. Under the same conditions, the specific rate of Red production was stimulated by increasing glucose levels, but markedly decreased by increased levels of phosphate.

Effect of post-induction nutrient feeding strategies on the production of bioadhesive protein in Escherichia coli.

The effect of post-induction nutrient feeding strategies on the production of bioadhesive protein using an IPTG inducible expression system in Escherichia coli was investigated. Cells were cultured in an exponential fed-batch mode to the OD600 of ca. 100 (48 gDCW/L) prior to induction. Six different post-induction nutrient feeding strategies (pH-stat, exponential, constant and linear change in feeding rate with three different slopes) were then applied, and bioadhesive protein production was examined. It was found that post-induction cell growth was independent of nutrient feeding rate. However, bioadhesive protein production was significantly affected by post-induction feeding strategies. Linearly changing post-induction feeding rate with a suitable slope allowed production of bioadhesive protein up to 5.3 g/L, which was higher than that obtained by the other post-induction feeding strategies.

Actinorhodin and undecylprodigiosin production in wild-type and relA mutant strains of Streptomyces coelicolor A3(2) grown in continuous culture.

The effects of growth rate and nutrient feed rate on the production of actinorhodin (Act) and undecylprodigiosin (Red) were determined in Streptomyces coelicolor A3(2) and in a congenic relA null-mutant known to be deficient in ppGpp synthesis and antibiotic production under conditions of nitrogen limitation. In the relA+ strain, Act production was inversely related to specific growth rate in continuous cultures limited by glucose, ammonium, or phosphate, while Red biosynthesis was optimal at 0.05 h-1 regardless of the specific nutrient limitation. Production of Act and Red in the relA mutant was lower than that of the parental strain, particularly under conditions of glucose- and ammonium-limitation, indicating an important and general role for ppGpp in determining the onset of the antibiotic biosynthesis under conditions of nutrient limitation. At constant growth rate, but with varying nutrient feed rates, the specific rate of Act production was adversely influenced by increasing levels of glucose, ammonium, and phosphate, with phosphate having the greatest inhibitory effect. Under the same conditions, the specific rate of Red production was stimulated by increasing glucose levels, but markedly decreased by increased levels of phosphate.

A New Methodology for Model Free Optimization of L-Lysine Biosynthesis

Economic viability of lysine, a process intensive in its use of raw materials, is dependent upon how efficiently these materials are used. This paper discusses the lysine optimization problem in fed-batch culture through optimization of initial media composition and control of nutrient feed rate. Previous strategies for solving these problems are discussed and their limitations are identified. An analytical approach based on metabolic flux balancing is suggested for optimizing intial media composition. Additionally, a new methodology for controlling the feed rate is proposed.

CONTINUOUS PROPIONIC ACID FERMENTATION OF HYDROLYZED COD (GADUS MORHUA) GURRY

Hydrolyzed cod gurry to which 4.0% glucose was added yielded a final pH of 4.5 after 96 h of incubation at 32C following inoculation with a 1.0% (v/v) inoculum (∼ 108 cells/mL) of a propionic acid producing bacterial culture. The use of a 90% (v/v) inoculum still resulted in a prolonged incubation time of 72 h to reach a pH of 4.5. A continuous propionic acid fermentation was successfully achieved for a period of 18 days with the addition of 2.5% to 4.0% glucose and a nutrient feed rate of 30 mL/h using a 500 mL fermentation vessel. A constant pH of 4.5 with an accompanying propionic acid concentration of 1.4% was maintained. Extrapolation of experimental data indicated that a glucose concentration of 2.2% is optimum.

High viable cell concentration fed‐batch cultures of hybridoma cells through on‐line nutrient feeding

A hybridoma cell line was cultivated in fed-batch cultures using a low-protein, serum-free medium. On-line oxygen uptake rate (OUR) measurement was used to adjust the nutrient feeding rate based on glucose consumption, which was estimated on-line using the stoichiometric relations between glucose and oxygen consumption. Through on-line control of the nutrient feeding rate, not only sufficient were supplied for cell growth and antibody production, but also the concentrations of glucose and other important nutrients such as amino acids were maintained at low levels during the cell growth phase. During the cultivation, cell metabolism changed from high lactate production and low oxygen consumption to low lactate production and high oxygen consumption. As a result the accumulation of lactate was reduced and the growth phase was extended. In comparison with the batch cultures, in which cells reached a concentration of approximately 2 x 10(6) cells/mL, a very high concentration of 1.36 x 10(7) cells/mL with a high cell viability (>90%) was achieved in the fed-batch culture. By considering the consumption of glucose and amino acids, as well as the production of cell mass, metabolites, and antibodies, a well-closed material balance was established. Our results demonstrate the value of coupling on-line OUR measurement and the stoichiometric relations for dynamic nutrient feeding in high cell concentration fed batch cultures.