Feeding strategies for E. coli fermentations demanding an enriched environment.

The addition of a carbon nutrient feed to a fed-batch cultivation is often not enough to obtain satisfactory growth and/or production. In some cases, an additional feed with for example supplementary amino acids or complex media is required. This work presents the development of feeding strategies where more than one feed is required and the knowledge of the growth requirements is low. Simulations and cultivations with E. coli are shown using the proposed feed controllers which are based on a probing control concept. The strategies work well and they can be used to shorten the process development phase considerably.

A cultivation technique for E. coli fed-batch cultivations operating close to the maximum oxygen transfer capacity of the reactor.

A cultivation strategy combining the advantages of temperature-limited fed-batch and probing feeding control is presented. The technique was evaluated in fed-batch cultivations with E. coli BL21(DE3) producing xylanase in a 3 liter bioreactor. A 20% increase in cell mass was achieved and the usual decrease in specific enzyme activity normally observed during the late production phase was diminished with the new technique. The method was further tested by growing E. coli W3110 in a larger bioreactor (50 l). It is a suitable cultivation technique when the O2 transfer capacity of the reactor is reached and it is desired to continue to produce the recombinant protein.

Probing control of glucose feeding in Vibrio cholerae cultivations.

Infection with Vibrio cholerae is a significant problem in many developing countries. Cultivation of V. cholerae is used in production of cholera toxin B subunit, which is a component in a cholera vaccine. Fed-batch cultivations with V. cholerae in defined media have been conducted and reproducible results were obtained. A probing feeding strategy developed by Akesson for Escherichia coli cultivations has been tested. The strategy is working as well for V. cholerae as for E. coli in minimizing the amount of acetic acid formed and avoiding anaerobic conditions. At 2 h after the feed start most of the acetic acid accumulated during the batch phase is consumed. The resulting feed rate tends to be the highest possible with respect to the constraints from cell metabolism and mass transfer, thus maximizing productivity in terms of biomass. A cell dry weight of 20-23 g/l is obtained after 12 h of feeding.

Production of heterologous thermostable glycoside hydrolases and the presence of host-cell proteases in substrate limited fed-batch cultures of Escherichia coli BL21(DE3).

Metabolic stress is a phenomenon often discussed in conjunction with recombinant protein production in Escherichia coli. This investigation shows how heterologous protein production and the presence of host cell proteases is related to: (1) Isopropyl-beta- D-thiogalactopyranoside (IPTG) induction, (2) cell-mass concentration at the time of induction, and (3) the presence of metabolites (glutamic acid or those from tryptone soy broth) during the post-induction phase of high cell density fed-batch cultivations. Two thermostable xylanase variants and one thermostable cellulase, all originating from Rhodothermus marinus, were expressed in E. coli strain BL21 (DE3). A three-fold difference in the specific activity of both xylanase variants [between 7,000 and 21,000 U/(g cell dry weight)], was observed under the different conditions tested. Upon induction at high cell-mass concentrations employing a nutrient feed devoid of the metabolites above, the specific activity of the xylanase variants, was initially higher but decreased 2-3 h into the post-induction phase and simultaneously protease activity was detected. Furthermore, protease activity was detected in all induced cultivations employing this nutrient feed, but was undetected in uninduced control cultivations (final cell-mass concentration of 40 g/l(-1)), as well as in induced cultivations employing metabolite-supplemented nutrient feeds. By contrast, maximum specific cellulase activity [between 700 and 900 U/(g cell dry weight)] remained relatively unaffected in all cases. The results demonstrate that detectable host cell proteases was not the primary reason for the decrease in post-induction activity observed under certain conditions, and possible causes for the differing production levels of heterologous proteins are discussed.

Avoiding acetate accumulation in Escherichia coli cultures using feedback control of glucose feeding.

An automated glucose feeding strategy that avoids acetate accumulation in cultivations of Escherichia coli is discussed. We have previously described how a probing technique makes it possible to detect and avoid overflow metabolism using a dissolved oxygen sensor. In this article these ideas are extended with a safety net that guarantees that aerobic conditions are maintained. The method is generally applicable, as no strain-specific information is needed and the only sensor required is a standard dissolved oxygen probe. It also gives the highest feed rate possible with respect to limitations from overflow metabolism and oxygen transfer, thus maximizing bioreactor productivity. The strategy was implemented on three different laboratory-scale platforms and fed-batch cultivations under different operating conditions were performed with three recombinant strains, E. coli K-12 UL635, E. coli BL21(DE3), and E. coli K-12 UL634. In spite of disturbances from antifoam and induction of recombinant protein production, the method reproducibly gave low concentrations of acetate and glucose. The ability to obtain favorable cultivation conditions independently of strain and operating conditions makes the presented strategy a useful tool, especially in situations where it is important to get good results on the first attempt.

On-line detection of acetate formation in Escherichia coli cultures using dissolved oxygen responses to feed transients.

Recombinant protein production in Escherichia coli can be significantly reduced by acetate accumulation. It is demonstrated that acetate production can be detected on-line with a standard dissolved oxygen sensor by superimposing short pulses to the substrate feed rate. Assuming that acetate formation is linked to a respiratory limitation, a model for dissolved oxygen responses to transients in substrate feed rate is derived. The model predicts a clear change in the character of the transient response when acetate formation starts. The predicted effect was verified in fed-batch cultivations of E. coli TOPP1 and E. coli BL21(DE3), both before and after induction of recombinant protein production. It was also observed that the critical specific glucose uptake rate, at which acetate formation starts, was significantly decreased after induction. On-line detection of acetate formation with a standard sensor opens up new possibilities for feedback control of substrate feeding.

Control of an ethanol fermentation carried out with alginate entrapped Saccharomyces cerevisiae.

Process control of different reactor models for continuous production of ethanol from sucrose with immobilized yeast has been studied. An enzyme thermistor with immobilized invertase recorded the concentration of sucrose continuously. Ethanol was recorded by a membrane gas sensor with a SnO(2) semiconductor used as detector. A process computer controlled the substrate feed to keep substrate as well as ethanol concentration at preset values by using algorithms of varying complexity. It was thereby demonstrated that PID regulators as well as more advanced algorithms (Otto-Smith regulator, state feedback from a Kalman filter, and cascade control) are useful alternatives to maintain a constant concentration in the fermentor effluents. The time required for the system to return to predetermined conditions after various kinds of disturbances has been especially studied. It was shown that the more advanced regulator used the shorter time.

Hepatic uptake of synthetic human gastrin I (1-17) in humans.

The hepatic uptake of unlabeled synthetic human gastrin I (1-17) was determined in four unanesthetized patients who did not have hepatic or gastroduodenal disease. The hormone was given by brief infusion and at rates producing concentrations of immunoreactive gastrin within the physiologic range. We estimated the first-pass fractional hepatic uptake by comparing the results after a portal and a peripheral infusion in each patient. It was -0.01 +/- 0.14 (mean +/- SD), demonstrating lack of hepatic uptake of gastrin I (1-17) in humans.

Disappearance of insulin in man: variation with the plasma insulin level.

Clearance rates of unlabelled insulin were studied in 45 unanaesthetized non-diabetic humans. The clearance rate, as well as the pancreatic secretion rate, of endogenous insulin was estimated from steady-state concentrations in portal and arterial blood. The clearance rate of exogenous insulin was determined after brief intraportal infusion. In the basal fasting state, the endogenous plasma insulin level varied as closely with the clearance of endogenous insulin as with the rate of pancreatic secretion. During elevation of insulin by glucose infusion, it varied predominantly with the rate of insulin secretion. Clearance of exogenous insulin did not vary with the pre-test endogenous insulin level. The clearance of endogenous insulin increased form 11 ml . min-1 . kg-1 in the basal fasting state to 17 ml . min-1. kg-1 during glucose infusion. Clearance of exogenous insulin fell progressively with increasing dose, form 35 (8 mU/kg) to 14 (43 mU/kg) ml . min-1 . kg-1 at normoglycaemia and 23 (8 mU/kg) to 17 (34 mU/kg) ml . min-1 . kg-1 at hyperglycaemia. The clearance of endogenous insulin was lower than that of exogenous insulin at normoglycaemia, but of similar size during glucose infusion. It is concluded that variation in clearance rate is partly responsible for variation in plasma insulin concentration, particularly in the basal fasting state, and that the clearance rate is lower in the basal state than otherwise. To some extent, the low clearance values for endogenous insulin in the basal state may reflect poor specificity of the insulin radioimmunoassay.