The design of controllers for batch bioreactors

The implementation of control algorithms to batch bioreactors is often complicated by variations in process dynamics that occur during the course of fermentation. Such a wide operating range often renders the performance of fixed gain proportional-integral-differential (PID) controllers unsatisfactory. In this work, detailed studies on the control of batch fermentations are per formed. Two simple controller designs are presented with the intent to compensate for changing process dynamics. One design incorporates the concepts of static feedforward-feedback control. While this technique produces tighter control than feedback alone, it is not as successful as a controller based on gain scheduling. The gain-scheduling controller, a subclass of adaptive controllers, uses the oxygen uptake rate as an auxiliary variable to fine-tune the PID controller parameters. The control of oxygen tension in the bioreactor is used as a vehicle to convey the proposed ideas, analyses, and results. Simulation experiments indicate significant improvement in controller performance can be achieved by both of the proposed approaches even in the presence of measurement noise.     

Model‐based optimization and pO2 control of fed‐batch Escherichia coli and Saccharomyces cerevisiae cultivation processes

A model‐based approach for optimization and cascade control of dissolved oxygen partial pressure (pO₂) and maximization of biomass in fed‐batch cultivations is presented. The procedure is based on the off‐line model‐based optimization of the optimal feeding rate profiles and the subsequent automatic pO₂ control using a proposed cascade control technique. During the model‐based optimization of the process, feeding rate profiles are optimized with respect to the imposed technological constraints (initial and maximal cultivation volume, cultivation time, feeding rate range, maximal oxygen transfer rate and pO₂ level). The cascade pO₂ control is implemented using activation of cascades for agitation, oxygen enrichment, and correction of the preoptimized feeding rate profiles. The proposed approach is investigated in two typical fed‐batch processes with Escherichia coli and Saccharomyces cerevisiae. The obtained results show that it was possible to achieve sufficiently high biomass levels with respect to the given technological constraints and to improve controllability of the investigated processes.     

The effect of oxygen partial pressure in bioreactors on cell proliferation and subsequent differentiation of somatic embryos of Cyclamen persicum

The effect of the relative oxygen partial pressure (pO₂) in bioreactors on cell proliferation and subsequent differentiation of somatic embryos from suspension cultures of Cyclamen persicum Mill. was investigated. The growth rate of cell line 3738-VIII in growth-regulator containing medium in bioreactors at 5% pO₂ was slightly reduced in comparison to 10% and 20% pO₂. Cultures growing at 40% pO₂ had a lower growth rate, a markedly reduced cell viability and showed a decrease of the medium pH to 3.5. Because a pH-control with a setpoint of 3.3 caused cell death within 4 days, it was assumed, that the reason for the poor cell proliferation and viability in the cultures at 40% pO₂ was an effect of medium acidification rather than of the high O₂ partial pressure. A significantly higher number of germinating embryos was obtained from the cultures grown at 40% pO₂ than from those grown in flasks or in bioreactors at 5%, 10% and 20% pO₂. These results were specific for cell line 3738-VIII. Another cell line, 3736-12, did not show marked differences in cell proliferation, viability, pH or subsequent regeneration of somatic embryos when grown at different O₂ partial pressures. This revised version was published online in June 2006 with corrections to the Cover Date.     

Saccade control in a simulated robot camera-head system: neural net architectures for efficient learning of inverse kinematics

The high speed of saccades means that they cannot be guided by visual feedback, so that any saccadic control system must know in advance the correct output signals to fixate a particular retinal position. To investigate neural-net architectures for learning this inverse-kinematics problem we simulated a 4 deg-of-freedom robot camera-head system, in which the head could pan and tilt and the cameras pan and verge. The main findings were: (1) Linear nets, multilayer perceptrons (MLPs) trained by backpropagation, and cerebellar model arithmetic computers (CMACs) all learnt rapidly to 5–10% accuracy when given perfect error feedback. (2) For additional accuracy (down to 2%) two-layer nets learnt much faster than a single MLP or CMAC: the best combination tried was to have a CMAC learn the errors of a trained linear net. (3) Imperfect error signals were provided by a crude controller whose output was simply proportional to retinal input in the relevant axis, thereby providing a mechanism for (a) controlling the camera-head system when the feedforward neural net controller was wrong or inoperative, and (b) converting sensory error signals into motor error signals as required in supervised learning. It proved possible to train neural-net controllers using these imperfect error signals over a range of learning rates and crude-controller gains. These results suggest that appropriate neural-net architectures can provide practical, accurate and robust adaptive control for saccadic movements. In addition, the arrangement of a crude controller teaching a sophisticated one may be similar to that used by the primate saccadic system, with brainstem circuitry teaching the cerebellum.     

Glyoxylate decreases the oxygen sensitivity of nitrogenase activity and photosynthesis in the cyanobacterium Anabaena cylindrica

Raising the pO₂ reduced nitrogenase activity (C₂H₂ reduction) of Anabaena cylindrica for both glyoxylate-treated (5 mM) and untreated cells. The stimulation caused by glyoxylate, however, increased with increases of pO₂ from 2 to 99 kPa. As the pO₂ increased the net CO₂ fixation was lowered (Warburg effect) while the CO₂ compensation point increased. Glyoxylate partly relieved this sensitivity of net photosynthesis to oxygen and reduced the compensation point considerably. The cells used were preincubated in the dark to exhaust photosynthetic pools. A more pronounced reduction in sensitivity of nitrogenase to oxygen for glyoxylate-treated cells was evident when a preincubation in air with reduced pCO₂ (13 l l^-1) was used. This was, however, not evident until after a 10-h incubation in air. Before this point 2 kPa O₂ sustained the highest nitrogenase activity. Addition of 0.5 and 5 mM of HCO ₃ ^- to Anabaena cultures preincubated at low CO₂ levels (29 l l^-1) abolished the stimulatory effect of glyoxylate on the nitrogenase. Thus, the results sustain the suggestion that glyoxylate may act as an inhibitor of photorespiratory activities in cyanobacteria and can be used as a means of increasing their nitrogen and CO₂ fixation capacities.Abbreviation RuBP ribulose 1,5-bisphosphate     

The snowball effect in fed-batch bioreactions

The bioreactor will play an important role in future biological manufacturing. For economic profit, important profiles of the feed rate in fed-batch cultures have been discussed. Unfortunately, the optimal feed rate is less robust. In these studies there exists the snowball effect in a substrate-inhibited bioprocess, in which substrate is accumulated due to uncertain parameters in the model or feed-rate error. The snowball effect also exists in multi-substrate-limited processes. In further studies, the interaction between the substrates has been higher in essential substrates than in growth-enhancing substrates. In a typical fed-batch bioreactor, the amount of the product can be reduced to 1% or less when the snowball effect arises. A new control structure, i.e., an off-line optimized feedforward controller added to a gain-scheduling PI(2)D feedback controller, is proposed to eliminate the troublesome snowball effect. The proposed control strategy recovers the yield up to 95%. Moreover, the robustness of the proposed control structure is demonstrated by simulation.     

Regulatory O2 tensions for the synthesis of fermentation products in Escherichia coli and relation to aerobic respiration

In an oxystat, the synthesis of the fermentation products formate, acetate, ethanol, lactate, and succinate of Escherichia coli was studied as a function of the O₂ tension (pO₂) in the medium. The pO₂ values that gave rise to half-maximal synthesis of the products (pO_0.5) were 0.2–0.4 mbar for ethanol, acetate, and succinate, and 1 mbar for formate. The pO_0.5 for the expression of the adhE gene encoding alcohol dehydrogenase was approximately 0.8 mbar. Thus, the pO₂ for the onset of fermentation was distinctly lower than that for anaerobic respiration (pO_0.5≤ 5 mbar), which was determined earlier. An essential role for quinol oxidase bd in microaerobic growth was demonstrated. A mutant deficient for quinol oxidase bd produced lactate as a fermentation product during growth at microoxic conditions (approximately 10 mbar O₂), in contrast to the wild-type or a quinol-oxidase-bo-deficient strain. In the presence of nitrate, the amount of lactate was largely decreased. Therefore, under microoxic conditions, the pO₂ appears to be too high for (mixed acid) fermentation to function and too low for aerobic respiration by quinol oxidase bo. Received: 7 February 1997 / Accepted: 2 May 1997     

Leaf gas films of Spartina anglica enhance rhizome and root oxygen during tidal submergence

Gas films on hydrophobic surfaces of leaves of some wetland plants can improve O₂ and CO₂ exchange when completely submerged during floods. Here we investigated the in situ aeration of rhizomes of cordgrass (Spartina anglica) during natural tidal submergence, with focus on the role of leaf gas films on underwater gas exchange. Underwater net photosynthesis was also studied in controlled laboratory experiments. In field experiments, O₂ microelectrodes were inserted into rhizomes and pO₂ measured throughout two tidal submergence events; one during daylight and one during night‐time. Plants had leaf gas films intact or removed. Rhizome pO₂ dropped significantly during complete submergence and most severely during night. Leaf gas films: (1) enhanced underwater photosynthesis and pO₂ in rhizomes remained above 10 kPa during submergence in light; and (2) facilitated O₂ entry from the water into leaves so that rhizome pO₂ was about 5 kPa during darkness. This study is the first in situ demonstration of the beneficial effects of leaf gas films on internal aeration in a submerged wetland plant. Leaf gas films likely contribute to submergence tolerance of S. anglica and this feature is expected to also benefit other wetland plant species when submerged.     

Nodule structure and nitrogenase activity ofCoriaria arborea in response to varying pO2

When excised root nodules ofCoriaria arborea are assayed for nitrogenase activity at various pO₂ they show a broad optimum between 20 and 40 kPa O₂, with some evidence for adaptation. Continuous flow assays of nodulated root systems of intact plants indicate that Coriaria shows an acetylene induced decline in nitrogenase activity. When root systems were subject to step changes in pO₂ nitrogenase activity responded with a steep decline followed by a slower rise in activity both at lower and higher than ambient pO₂. Thus Coriaria nodules are able to adapt rapidly to oxygen levels well above and well below ambient. Measurement of nodule diffusion resistance showed that the adaptation is accompanied by rapid increase in resistance at above ambient pO₂ and decrease in resistance at below ambient pO₂. Plants grown with root systems at pO₂ from 5–40 kPa O₂ did not differ in growth or nodulation. The anatomy of Coriaria nodules shows they have a dense periderm which encircles the nodule and also closely invests the infected zone. The periderm is both thicker and more heavily suberised in nodules grown at high pO₂ than at low pO₂. Vacuum infiltration of India ink indicates that oxygen diffusion is entirely through the lenticel and via a small gap adjacent to the stele.     

Constant dissolved oxygen concentrations in cephalosporin C fermentation: Applicability of different controllers and effect on fermentation parameters

Summary The behaviour and applicability of several controllers for maintaining a constant dissolved oxygen concentration (DO) during the cephalosporin C production with Cephalosporium acremonium in a laboratory fermentor is described. The process controllers were realized on a MC 68000 based process computer using the real-time language PEARL. The discrete signum integral controller showed the best control action. In addition some derived fermentation data were calculated on-line by the process computer.The results obtained by comparison of fermentations carried out at DO between 10% and 40% saturation during ideophase indicate that high DO leads to a high specific production rate for cephalosporin C and a low specific production rate for penicillin N and vice versa. In the range of DO investigated the production of deacetyl and deacetoxy cephalosporin C is not affected by DO. A direct correlation between DO and the yield coefficients Y _P/S and Y _P/X could be established. The yield coefficient Y _P/O for cephalosporin C is constant in the DO range from 10%–40%.Dedicated to Prof. Dr. H. J. Rehm on the occasion of his 60th birthday     

Oxygen limitation of N2 fixation in stem-girdled and nitrate-treated soybean

The effects of increasing rhizosphere pO₂on nitrogenase activity and nodule resistance to O₂diffusion were investigated in soybean plants [Glycine max (L.) Merr. cv. Harosoy 63] in which nitrogenase (EC 1.7.99.2) activities were inhibited by (a) removal of the phloem tissue at the base of the stem (stem girdling), (b) exposure of roots to 10 mM NO₃over 5 days (NO₃-treated), or (c) partial inactivation of nitrogenase activity by an exposure of nodulated roots to 100 kPa O₂(O₂-inhibitcd). In control plants and in plants which had been treated with 100 kPa O₂, increasing rhizosphere O₂concentrations in 10 kPa increments from 20 to 70 kPa did not alter the steady-state nitrogenase activity. In contrast, in plants in which nitrogenase activities were depressed by stem girdling or by exposure to NO₃, increasing rhizosphere pO₂resulted in a recovery of 57 or 67%, respectively, of the initial, depressed rates of nitrogenase activity. This suggests that the nitrogenase activity of stem-girdled and NO₃-treated soybeans was O₂-limited. For each treatment, theoretical resistance values for O₂diffusion into nodules were estimated from measured rates of CO₂exchange, assuming a respiratory quotient of 1.1 and 0 kPa of O₂in the infected cells. At an external partial pressure of 20 kPa O₂, the stem-girdled and NO₃^--treated plants displayed resistance values which were 4 to 8.6 times higher than those in the nodules of the control plants. In control and O₂-inhibited plants, increases in pO₂from 20 to 70 kPa in 10 kPa increments resulted in a 2.5- to 3.9-fold increase in diffusion resistance to O₂, and had little effect on either respiration or nitrogenase activity. In contrast, in stem-girdled and NO₃^--treated plants, increases in external pO₂had little effect on diffusion resistance to O₂, but resulted in a 2.3- to 3.2-fold increase in nodule respiration and nitrogenase activity. These results are consistent with stem-girdling and NO₃^--inhibition treatments limiting phloem supply to nodules causing an increase in diffusion resistance to O₂at 20 kPa and an apparent insensitivity of diffusion resistance to increases in external pO₂.     

Survival and function of mesenchymal stem cells (MSCs) depend on glucose to overcome exposure to long‐term,severe and continuous hypoxia

Use of mesenchymal stem cells (MSCs) has emerged as a potential new treatment for various diseases but has generated marginally successful results. A consistent finding of most studies is massive death of transplanted cells. The present study examined the respective roles of glucose and continuous severe hypoxia on MSC viability and function with respect to bone tissue engineering. We hereby demonstrate for the first time that MSCs survive exposure to long‐term (12 days), severe (pO₂ < 1.5 mmHg) hypoxia, provided glucose is available. To this end, an in vitro model that mimics the hypoxic environment and cell‐driven metabolic changes encountered by grafted sheep cells was established. In this model, the hallmarks of hypoxia (low pO₂, hypoxia inducible factor‐1α expression and anaerobic metabolism) were present. When conditions switched from hypoxic (low pO₂) to ischemic (low pO₂ and glucose depletion), MSCs exhibited shrinking, decreased cell viability and ATP content due to complete exhaustion of glucose at day 6; these results provided evidence that ischemia led to the observed massive cell death. Moreover, MSCs exposed to severe, continuous hypoxia, but without any glucose shortage, remained viable and maintained both their in vitro proliferative ability after simulation with blood reperfusion at day 12 and their in vivo osteogenic ability. These findings challenge the traditional view according to which severe hypoxia per se is responsible for the massive MSC death observed upon transplantation of these cells and provide evidence that MSCs are able to withstand exposure to severe, continuous hypoxia provided that a glucose supply is available.     

Does the agitation rate and/or oxygen saturation influence exopolysaccharide production by Aureobasidium pullulans in batch culture?

  When Aureobasidium pullulans was grown at a number of agitation rates under batch conditions, exopolysaccharide yields were dramatically reduced at high rates i.e. at least 750 rpm. Investigations with gas blending, which allowed pO₂ manipulation and control independently of the agitation rate, showed that this yield reduction was due solely to the high pO₂ levels that occurred at these agitation rates. Thus, polysaccharide production at 1000 rpm could be elevated by maintaining the pO₂ at a low level during the initial phase of the fermentation. However, both the timing of the pO₂ decrease and the level at which it was maintained were crucial for obtaining yields at 1000 rpm, similar to those observed at low agitation rates. Received: 29 February 1996 / Received revision: 11 July 1996 / Accepted: 15 July 1996     

Contextual cues are not unique for motor learning: Task-dependant switching of feedback controllers

The separation of distinct motor memories by contextual cues is a well known and well studied phenomenon of feedforward human motor control. However, there is no clear evidence of such context-induced separation in feedback control. Here we test both experimentally and computationally if context-dependent switching of feedback controllers is possible in the human motor system. Specifically, we probe visuomotor feedback responses of our human participants in two different tasks—stop and hit—and under two different schedules. The first, blocked schedule, is used to measure the behaviour of stop and hit controllers in isolation, showing that it can only be described by two independent controllers with two different sets of control gains. The second, mixed schedule, is then used to compare how such behaviour evolves when participants regularly switch from one task to the other. Our results support our hypothesis that there is contextual switching of feedback controllers, further extending the accumulating evidence of shared features between feedforward and feedback control.     

Effect of oxygen pressure on synthesis and export of nitrogenous solutes by nodules of cowpea

Nodules of cowpea plants (Vigna unguiculata (L.) Walp. cv. Vita 3 :Bradyrhizobium CB756) cultured for periods of 23 d with their root systems maintained in atmospheres containing a range of partial pressures of O₂ (pO₂; 1–80%, v/v, in N₂) formed and exported ureides (allantoin and allantoic acid) as the major products of fixation at all pO₂ tested. In sub-ambient pO₂ (1 and 2.5%) nodules contained specific activities of uricase (urate: O₂ oxidoreductase; EC 1.7.3.3) and allantoinase (allantoin hydrolyase; EC 3.5.2.5) as much as sevenfold higher than in those from air. On a cell basis, uninfected cells in nodules from 1% O₂ contained around five times the level of uricase. Except for NAD: glutamate synthase (EC 1.4.1.14), which was reduced in sub-ambient O₂, the activities of other enzymes of ureide synthesis were relatively unaffected by pO₂. Short-term effects of pO₂ on assimilation of fixed nitrogen were measured in nodules of air-grown plants exposed to subambient pO₂ (1, 2.5 or 5%, v/v in N₂) and¹⁵N₂. Despite a fall in total¹⁵N₂ fixation, ureide synthesis and export was maintained at a high level except in 1% O₂ where formation was halved. The data indicate that in addition to the structural and diffusional adaptations of cowpea nodules which allow the balance between O₂ supply and demand to be maintained over a wide range of pO₂, nodules also show evidence of biochemical adaptations which maintain and enhance normal pathways for the assimilation of fixed nitrogen. This work was supported by a grant from the Australian Research Council (to C.A.A.) and an Australian Development Assistance Bureau postgraduate fellowship (to F.D.D.).     

The cultivation of animal cells at controlled dissolved oxygen partial pressure

A 3-liter culture vessel has been developed for the growth of animal cells in suspension at controlled pH and dissolved oxygen partial pressure (pO₂). The culture technique allows metabolically produced CO₂ to be measured; provision can be made to control the dissolved CO₂ partial pressure. In cultures containing a low serum concentration, gas sparging to control pO₂ was found to cause cell damage. This could be prevented by increasing the serum concentration to 10%, or by adding 0.02% of the surface-active polymer Pluronic F68. The growth of mouse LS cells in batch culture without pO₂ control was found to be limited by the availability of oxygen. Maximum viable cell populations were obtained when dissolved pO₂ was controlled at values within the range 40–100 mm Hg.     

Accurate control of oxygen level in cells during culture on silicone rubber membranes with application to stem cell differentiation

Oxygen level in mammalian cell culture is often controlled by placing culture vessels in humidified incubators with a defined gas phase partial pressure of oxygen (pO_2gas). Because the cells are consuming oxygen supplied by diffusion, a difference between pO_2gas and that experienced by the cells (pO_2cell) arises, which is maximal when cells are cultured in vessels with little or no oxygen permeability. Here, we demonstrate theoretically that highly oxygen‐permeable silicone rubber membranes can be used to control pO_2cell during culture of cells in monolayers and aggregates much more accurately and can achieve more rapid transient response following a disturbance than on polystyrene and fluorinated ethylene‐propylene copolymer membranes. Cell attachment on silicone rubber was achieved by physical adsorption of fibronectin or Matrigel. We use these membranes for the differentiation of mouse embryonic stem cells to cardiomyocytes and compare the results with culture on polystyrene or on silicone rubber on top of polystyrene. The fraction of cells that are cardiomyocyte‐like increases with decreasing pO₂ only when using oxygen‐permeable silicone membrane‐based dishs, which contract on silicone rubber but not polystyrene. The high permeability of silicone rubber results in pO_2cell being equal to pO_2gas at the tissue‐membrane interface. This, together with geometric information from histological sections, facilitates development of a model from which the pO₂ distribution within the resulting aggregates is computed. Silicone rubber membranes have significant advantages over polystyrene in controlling pO_2cell, and these results suggest they are a valuable tool for investigating pO₂ effects in many applications, such as stem cell differentiation. © 2009 American Institute of Chemical Engineers Biotechnol. Prog., 2010     

Influence of salinity and abscisic acid on the O2 uptake by N2-fixing nodules of common bean

The effects of NaCl and ABA on the respiration of N₂-fixing nodules were analysed in common bean (Phaseolus vulgaris) inoculated with Rhizobium tropici the reference strain CIAT899. Shoot and nodule growth was more inhibited by NaCl than root growth. The O₂ uptake by nodulated roots at 21 kPa O₂ was significantly inhibited by salinity. Raising pO₂ stimulated nodule respiration more under NaCl treatment than for the control, although it did not compensate totally for the inhibitory effect of NaCl. Short NaCl application was less destructive than long term application. Also, the external application of ABA inhibited nodule respiration, and this inhibition was partly compensated by raising pO₂.     

Utilization of simple phenolics for dinitrogen fixation by soil diazotrophic bacteria

Summary A microaerobic diazotrophic bacterium tentatively identified as aPseudomonas species was isolated from a forest soil. Its nitrogenase (C₂H₂ reduction) activity in liquid medium was significantly supported by phenolic compounds when compared with glucose-, mannitol- or malate-supported activity. The utilization of phenolics was dependent on substrate induction and the appropriate oxygen concentration. At a pO₂ of 0.05 protocatechuate was a better carbon source for N₂ fixation than glucose. In the case ofLignobacter protocatechuate was a better carbon source for N₂ fixation than glucose at pO₂ 0.2 but not at pO₂ 0.05. It is suggested that certain monomeric phenols can support nitrogenase activities in many carbon-limited soil environments.Contribution No. 1484 from the Chemistry and Biology Research Institute, Agriculture Canada, Ottawa, Canada.     

Physiological description of multivariate interdependencies between process parameters,morphology and physiology during fed‐batch penicillin production

Optimization of productivity and economics of industrial bioprocesses requires characterization of interdependencies between process parameters and process performance. In the case of penicillin production, as in other processes, process performance is often closely interlinked with the physiology and morphology of the organism used for production. This study presents a systematic approach to efficiently characterize the physiological effects of multivariate interdependencies between bioprocess design parameters (spore inoculum concentration, pO₂ control level and substrate feed rate), morphology, and physiology. Method development and application was performed using the industrial model process of penicillin production. Applying traditional, statistical bioprocess analysis, multivariate correlations of raw bioprocess design parameters (high spore inoculum concentration, low pO₂ control as well as reduced glucose feeding) and pellet morphology were identified. A major drawback of raw design parameter correlation models; however, is the lack of transferability across different process scales and regimes. In this context, morphological and physiological bioprocess modeling based on scalable physiological parameters is introduced. In this study, raw parameter effects on pellet morphology were efficiently summarized by the physiological parameter of the biomass yield per substrate. Finally, for the first time to our knowledge, the specific growth rate per spore was described as time‐independent determinant for switching from pellet to disperse growth during penicillin production and thus introduced as a novel, scalable key process parameter for pellet morphology and process performance. © 2014 American Institute of Chemical Engineers Biotechnol. Prog., 30:689–699, 2014