The scale up of mycelial shake flask fermentations: A case study of gamma linolenic acid production byMucor hiemalis IRL 51

Summary Production of gamma linolenic acid (GLA) by the filamentous fungusMucor hiemalis IRL 51 was studied in both shake flask culture and in a 10-L stirred tank fermenter. This study was conducted to assess how the results from shake flask media screening trials compared to those obtained in a 10-L stirred tank fermenter, which is assumed to be more representative of an industrial system. The results show that the biological performance in 10-L fermenters is usually the same as that in shake flask culture. There were some inconsistencies which could possibly be attributed to scale, but no large differences were systematically seen. These results show that for this filamentous fungus, shake flask culture provides a quick and inexpensive way of optimizing medium composition.     

New milliliter‐scale stirred tank bioreactors for the cultivation of mycelium forming microorganisms

A novel milliliter‐scale stirred tank bioreactor was developed for the cultivation of mycelium forming microorganisms on a 10 milliliter‐scale. A newly designed one‐sided paddle impeller is driven magnetically and rotates freely on an axis in an unbaffled reaction vessel made of polystyrene. A rotating lamella is formed which spreads out along the reactor wall. Thus an enhanced surface‐to‐volume ratio of the liquid phase is generated where oxygen is introduced via surface aeration. Volumetric oxygen transfer coefficients (k_La) > 0.15 s^?1 were measured. The fast moving liquid lamella efficiently prevents wall growth and foaming. Mean power consumption and maximum local energy dissipation were measured as function of operating conditions in the milliliter‐scale stirred tank bioreactor (V = 10 mL) and compared to a standard laboratory‐scale stirred tank bioreactor with six‐bladed Rushton turbines (V = 2,000 mL). Mean power consumption increases with increasing impeller speed and shows the same characteristics and values on both scales. The maximum local energy dissipation of the milliliter‐scale stirred tank bioreactor was reduced compared to the laboratory‐scale at the same mean volumetric power input. Hence the milliliter impeller distributes power more uniformly in the reaction medium. Based on these data a reliable and robust scale‐up of fermentation processes is possible. This was demonstrated with the cultivation of the actinomycete Streptomyces tendae on both scales. It was shown that the process performances were equivalent with regard to biomass concentration, mannitol consumption and production of the pharmaceutical relevant fungicide nikkomycin Z up to a process time of 120 h. A high parallel reproducibility was observed on the milliliter‐scale (standard deviation < 8%) with up to 48 stirred tank bioreactors operated in a magnetic inductive drive. Rheological behavior of the culture broth was measured and showed a highly viscous shear‐thinning non‐Newtonian behavior. The newly developed one‐sided paddle impellers operated in unbaffled reactors on a 10 milliliter‐scale with a magnetic inductive drive for up to 48 parallel bioreactors allows for the first time the parallel bioprocess development with mycelium forming microorganisms. This is especially important since these kinds of cultivations normally exhibit process times of 100 h and more. Thus the operation of parallel stirred tank reactors will have the potential to reduce process development times drastically. Biotechnol. Bioeng. 2010; 106: 443–451. © 2010 Wiley Periodicals, Inc.     

Protein crystallization in stirred systems—scale‐up via the maximum local energy dissipation

Macromolecular bioproducts like therapeutic proteins have usually been crystallized with µL‐scale vapor diffusion experiments for structure determination by X‐ray diffraction. Little systematic know‐how exists for technical‐scale protein crystallization in stirred vessels. In this study, the Fab‐fragment of the therapeutic antibody Canakinumab was successfully crystallized in a stirred‐tank reactor on a 6 mL‐scale. A four times faster onset of crystallization of the Fab‐fragment was observed compared to the non‐agitated 10 µL‐scale. Further studies on a liter‐scale with lysozyme confirmed this effect. A 10 times faster onset of crystallization was observed in this case at an optimum stirrer speed. Commonly suggested scale‐up criteria (i.e., minimum stirrer speed to keep the protein crystals in suspension or constant impeller tip speed) were shown not to be successful. Therefore, the criterion of constant maximum local energy dissipation was applied for scale‐up of the stirred crystallization process for the first time. The maximum local energy dissipation was estimated by measuring the drop size distribution of an oil/surfactant/water emulsion in stirred‐tank reactors on a 6 mL‐, 100 mL‐, and 1 L‐scale. A comparable crystallization behavior was achieved in all stirred‐tank reactors when the maximum local energy dissipation was kept constant for scale‐up. A maximum local energy dissipation of 2.2 W kg^?1 was identified to be the optimum for lysozyme crystallization at all scales under study. Biotechnol. Bioeng. 2013; 110: 1956–1963. © 2013 Wiley Periodicals, Inc.     

Mineral oil addition as a means of foam control for plant cell cultures in stirred tank fermenters

Summary The addition of light mineral oil was found to be effective in preventing foaming during plant cell cultivation in stirred tank fermenters without any adverse effect on cell growth. The optimum amount of oil is about 5% (by volume) of the total liquid medium. The best time for the addition is after the lag period of the cell growth cycle, which is about 2 days after cultivation.     

A comparison of laboratory and pilot-scale fermentations in winemaking conditions

We investigated the influence of the fermenter size on alcoholic fermentation. Experiments were carried out at pilot scale, in 100-L fermenters, and at laboratory scale, in stirred and static 1-L fermenters. Two musts, Grenache blanc and Sauvignon, were fermented with and without the addition of solid particles from grape musts. Highly clarified must fermentation kinetics was strongly affected by the scale of the experiment, with slower fermentation occurring in the 100-L fermenter. Alcohol, ester, and thiol synthesis in clarified sauvignon must fermentation was also strongly correlated with the fermentation scale. Addition of solid particles from grape tended to reduce the effects on kinetics associated with increasing the scale of the fermentation, by increasing the maximum rate of CO₂ production, and by shortening the duration of fermentation. The addition of such particles also decreased the effects of scaling up the fermentation on the concentration of some volatile compounds, i.e., isoamyl acetate, ethyl octanoate, but did not decrease this effect for other compounds, such as isobutyl acetate, isobutanol, and 3-mercaptohexanol.     

Phytase production by Sporotrichum thermophile in a cost‐effective cane molasses medium in submerged fermentation and its application in bread

Aims: Phytase production by Sporotrichum thermophile in a cost‐effective cane molasses medium in submerged fermentation and its application in bread. Methods and Results: The production of phytase by a thermophilic mould S. thermophile was investigated using free and immobilized conidiospores in cane molasses medium in shake flasks, and stirred tank and air‐lift fermenters. Among surfactants tested, Tweens (Tween‐20, 40 and 80) and sodium oleate increased phytase accumulation, whereas SDS and Triton X‐100 inhibited the enzyme production. The mould produced phytase optimally at a_w 0·95, and it declined sharply below this a_w value. The enzyme production was comparable in air‐lift and stirred tank reactors with a marked reduction in fermentation time. Among the matrices tried, Ca‐alginate was the best for conidiospore immobilization, and fungus secreted sustained levels of enzyme titres over five cycles. The phytic acid in the dough was efficiently hydrolysed by the enzyme accompanied by the liberation of soluble phosphate in the bread. Conclusions: The phytase production by S. thermophile was enhanced in the presence of Tween‐80 in cane molasses medium. A peak in enzyme production was attained in 48 h in the fermenter when compared with that of 96 h in shake flasks. Ca‐alginate immobilized conidiospores germinated to produce fungal growth that secreted sustained levels of phytase over five cycles. The bread made with phytase contained reduced level of phytic acid and a high‐soluble phosphate. Significance and Impact of the Study: The phytase accumulation by S. thermophile was increased by the surfactants. The sustainability of enzyme production in stirred tank and air‐lift fermenters suggested the possibility for scaling up of phytase. The bread made with phytase contained low level of antinutrient, i.e. phytic acid.     

Comparison of manufacturing techniques for adenovirus production

We have compared three different production methods, which may be suitable for the large scale production of adenovirus vectors for human clinical trials. The procedures compared 293 cells adapted to suspension growth in serum-free medium in a stirred tank bioreactor, 293 cells on microcarriers in serum-containing medium in a stirred tank bioreactor, and 293 cells grown in standard tissue culture plasticware. With a given virus, yields varied between 2000 and 10,000 infectious units/cell. The stirred tank bioreactor routinely produced between 4000 and 7000 infectious units/cell when 293 cells were grown on microcarriers. The 293 cells adapted to suspension growth in serum-free medium in the same stirred tank bioreactor yielded between 2000 and 7000 infectious units/cell. Yields obtained from standard tissue culture plasticware were up to 10,000 infectious units/cell. Cell culture conditions were monitored for glucose consumption, lactate production, and ammonia accumulation. Glucose consumption and lactate accumulation correlated well with the cell growth parameters. Ammonia production does not appear to be significant. Based on virus yields, ease of operation and linear scalability, large-scale adenovirus production seems feasible using 293 cells (adapted to suspension/serum free medium or on microcarriers in serum containing medium) in a stirred tank bioreactor. This revised version was published online in July 2006 with corrections to the Cover Date.     

Degradation of n-valeric acid by alginate-entrapped Alcaligenes denitrificans

Summary Alcaligenes denitrificans was isolated from sewage sludge and showed a strong degradative ability towards volatile fatty acids. The organism was tested both as free cells and immobilised in calcium alginate, for the ability to degrade the sodium salt of a typical volatile fatty acid, valeric acid.In shake flask culture the immobilised cells could be used to fully degrade 18 mM valerate over ten 48 h runs before bead break up occurred. The use of beads in conventional stirred tank fermenters, and a bubble column reactor was also investigated, with a 50 ml bubble column containing 5 ml of beads giving the highest overall degradation rate of 1.8 mmol/h, for 40 h in a fed batch mode of operation.     

Process modeling of in situ-adsorption of a bacterial lipase

In situ adsorption, known as an in situ-roduct removal (ISPR) technique for low molecular mass bioproducts, was in this study applied to a bacterial exoenzyme proving that this method is also suitable for the separation of macromolecules like proteins. For this, adsorbent particles were added to growing cultures of Staphylococcus carnosus rec., therefore both production and adsorption occurred simultaneously in shaking flasks, stirred tank, or airlift bioreactor as the chosen types of fermenters. The exoenzyme lipase adsorbed rapidly and, after separating cells and adsorbents, desorbed in a packed bed column. Up to 85% of the produced lipase were recovered, fractions of these had been concentrated up to the factor 20 and purified up to a factor of 40 by the procedure. By using the airlift bioreactor an enhancement of biomass production was observed, but the necessity of the addition of an anti-foam reagent resulted in higher product losses in adsorption as well as in desorption. Production and adsorption kinetics have been modeled and applied to in situ-adsorption. The model was used to perform a parameter study in which the influence of biological and physical parameters as well as process parameters on discontinuous and continuous in situ-adsorption was investigated.     

Power consumption and maximum energy dissipation in a milliliter‐scale bioreactor

Mean power consumption and maximum local energy dissipation were measured as function of operating conditions of a milliliter‐scale stirred tank bioreactor (V = 12 mL) with a gas‐inducing impeller. A standard laboratory‐scale stirred tank bioreactor (V = 1,200 mL) with Rushton turbines was used as reference. The measured power characteristics (Newton number as function of Reynolds number) were the same on both scales. The changeover between laminar and turbulent flow regime was observed at a Reynolds number of 3,000 with the gas‐inducing stirrer on a milliliter‐scale. The Newton number (power number) in the turbulent flow regime was 3.3 on a milliliter‐scale, which is close to values reported for six‐blade Rushton turbines of standard bioreactors. Maximum local energy dissipation (ε_max) was measured using a clay/polymer flocculation system. The maximum local energy dissipation in the milliliter‐scale stirred tank bioreactor was reduced compared with the laboratory‐scale stirred tank at the same mean power input per unit mass (ε_ø), yielding ε_max/ε_ø ≈ 10 compared with ε_max/ε_ø ≈ 16. Hence, the milliliter‐scale stirred tank reactor distributes power more uniformly in the reaction medium. These results are in good agreement with literature data, where a decreasing ε_max/ε_ø with increasing ratio of impeller diameter to reactor diameter is found (d/D = 0.7 compared with d/D = 0.4). Based on these data, impeller speeds can now be easily adjusted to achieve the same maximum local energy dissipation at different scales. This enables a more reliable and robust scale‐up of bioprocesses from milliliter‐scale to liter‐scale reactors. © 2009 American Institute of Chemical Engineers Biotechnol. Prog., 2010     

Effect of oxygen supply on the suspension culture of genetically modified tobacco cells.

The effect of oxygen supply on the cultivation of the genetically modified tobacco cells and the formation of a foreign protein, beta-glucuronidase (GUS), was investigated in 250-mL Erlenmeyer flasks, a 5-L stirred tank fermenter, and a 7-L air-lift fermenter. The oxygen supply was varied by using different volumes of medium in the case of the 250-mL Erlenmeyer flask culture or by the different aeration rate in the case of the two types of fermenters tested. Higher oxygen supply stimulated cell growth and increased oxygen consumption rate, the level of phenolics, and GUS productions.     

Monoterpenoid oxindole alkaloid production by Uncaria tomentosa (Willd) D.C. cell suspension cultures in a stirred tank bioreactor

Cell growth, monoterpenoid oxindole alkaloid (MOA) production, and morphological properties of Uncaria tomentosa cell suspension cultures in a 2-L stirred tank bioreactor were investigated. U. tomentosa (cell line green Uth-3) was able to grow in a stirred tank at an impeller tip speed of 95 cm/s (agitation speed of 400 rpm), showing a maximum biomass yield of 11.9 +/- 0.6 g DW/L and a specific growth rate of 0.102 d(-1). U. tomentosa cells growing in a stirred tank achieved maximum volumetric and specific MOA concentration (467.7 +/- 40.0 microg/L, 44.6 +/- 5.2 microg/g DW) at 16 days of culture. MOA chemical profile of cell suspension cultures growing in a stirred tank resembled that of the plant. Depending on culture time, from the total MOA produced, 37-100% was found in the medium in the bioreactor culture. MOA concentration achieved in a stirred tank was up to 10-fold higher than that obtained in Erlenmeyer flasks (agitated at 110 rpm). In a stirred tank, average area of the single cells of U. tomentosa increased up to 4-fold, and elliptical form factor increased from 1.40 to 2.55, indicating enlargement of U. tomentosa single cells. This work presents the first report of U. tomentosa green cell suspension cultures that grow and produce MOA in a stirred tank bioreactor.     

Advances in clone selection using high‐throughput bioreactors

Effective clone selection is a crucial step toward developing a robust mammalian cell culture production platform. Currently, clone selection is done by culturing cells in well plates and picking the highest producers. Ideally, clone selection should be done in a stirred tank bioreactor as this would best replicate the eventual production environment. The actual number of clones selected for future evaluation in bioreactors at bench‐scale is limited by the scale‐up and operational costs involved. This study describes the application of miniaturized stirred high‐throughput bioreactors (35 mL working volume; HTBRs) with noninvasive optical sensors for clone screening and selection. We investigated a method for testing several subclones simultaneously in a stirred environment using our high throughput bioreactors (up to 12 clones per HTBR run) and compared it with a traditional well plate selection approach. Importantly, it was found that selecting clones solely based on results from stationary well plate cultures could result in the chance of missing higher producing clones. Our approach suggests that choosing a clone after analyzing its performance in a stirred bioreactor environment is an improved method for clone selection. © 2010 American Institute of Chemical Engineers Biotechnol. Prog., 2010     

Production of bacteriocin jenseniin G by Propionibacterium is pH sensitive

Bacteriocin jenseniin G is detected typically in 50-fold concentrated 10 d cultures of the producer, Propionibacterium thoenii ( jensenii ) P126 at low concentrations (32–64 AU ml⁻¹). The effect of pH on production was examined by growing the producer in sodium lactate broth, both statically and in stirred fermenters, with or without pH control. Activity was detected in unconcentrated static producer cultures at day 7 through to day 14. Maximal jenseniin G activity (21 AU ml⁻¹) was observed in concentrated supernates at day 9, remaining relatively constant through to day 14. Producer cultures grown in stirred fermenters without pH control yielded detectable activity in 50-fold concentrated supernates at day 3 and maximal activity at day 11. Producer growth in stirred fermenters at controlled pH yielded maximum production at pH 6·40 and maximum activity (160 AU ml⁻¹) at day 13. Jenseniin G activity at pH 6·40 represented a fivefold increase over previous reports. Medium pH was critical to jenseniin G production.     

Shear rate and microturbulence effects on the synthesis of proteases by Jacaratia mexicana cells cultured in a bubble column, airlift, and stirred tank bioreactors

Cysteine proteases from Jacaratia mexicana, an endemic Mexican plant, could compete in industrial applications with papain. Currently the only way to obtain these proteases is by extracting them from the wild plant. An alternative source of these enzymes is by J. mexicana suspension culture. In this work, this culture was carried out in airlift, bubble column and stirred tank bioreactors, and the effects of shear rate and microturbulence on cell growth, protein accumulation and proteolytic activity were determined. The shear rates in the stirred tank, bubble column and airlift bioreactors were 274 1/s, 13 1/s and 36 1/s respectively, and microturbulences (symbolized by λ, in units of μm) were 46, 79, and 77 μm, respectively. Protein levels and proteolytic activity were linearly correlated with both shear rate and microturbulence. A higher shear rate and a more intensive microturbulence occurred in the stirred tank, producing higher protein accumulation and higher proteolytic activity compared with those of the other two bioreactor systems. Higher shear rate and microturbulence had an elicitor effect on protease synthesis, because microturbulence in stirred tank bioreactors was lower than the average length of J. mexicana cells. Furthermore, cells in the stirred tank were smaller and thinner than those grown in shake flask, bubble column and airlift bioreactors. In summary, proteases were produced by J. mexicana cell cultures in a stirred tank under conditions of high shear rate and intensive microturbulence, which are similar to those which occur in industrial stirred tanks. These results encourage continuation of the process development for large scale production of these proteases by this technology.     

Large-scale preparation of recombinant platelet-derived growth factor AA secreted from recombinant baby hamster kidney cells

The short isoform of platelet-derived growth factor A (PDGF-A) was expressed in a mammalian host (BHK-21 cell). A cell line was obtained that secreted up to 0.3 micrograms/10(6) cells recombinant PDGF-A chain homodimer/day into the medium. For large-scale production of supernatant, cells were grown either in roller bottles or in 2.5-1 stirred tank fermenters. A simple two-step procedure was developed to purify recombinant PDGF-AA (rPDGF-AA). The first step was adsorption onto porous glass and the final step was reversed-phase high-performance liquid chromatography. The yield was 0.2 mg/l supernatant. A total amount of 20-30 mg pure rPDGF-AA may be obtained from a single fermenter run. Sequence analysis showed the correct amino terminus and no internal proteolytic cleavages. The specific activity was 5 ng/ml for mouse AKR-2B cells. [125I]rPDGF-AA had an affinity constant of approximately 0.5 nM to these cells and 25,000 binding sites were estimated/cell.     

High cell density cultivation of recombinant yeasts and bacteria under non-pressurized and pressurized conditions in stirred tank bioreactors

This study demonstrates the applicability of pressurized stirred tank bioreactors for oxygen transfer enhancement in aerobic cultivation processes. The specific power input and the reactor pressure was employed as process variable. As model organism Escherichia coli, Arxula adeninivorans, Saccharomyces cerevisiae and Corynebacterium glutamicum were cultivated to high cell densities. By applying specific power inputs of approx. 48kWm(-3) the oxygen transfer rate of a E. coli culture in the non-pressurized stirred tank bioreactor was lifted up to values of 0.51moll(-1)h(-1). When a reactor pressure up to 10bar was applied, the oxygen transfer rate of a pressurized stirred tank bioreactor was lifted up to values of 0.89moll(-1)h(-1). The non-pressurized stirred tank bioreactor was able to support non-oxygen limited growth of cell densities of more than 40gl(-1) cell dry weight (CDW) of E. coli, whereas the pressurized stirred tank bioreactor was able to support non-oxygen limited growth of cell densities up to 225gl(-1) CDW of A. adeninivorans, 89gl(-1) CDW of S. cerevisiae, 226gl(-1) CDW of C. glutamicum and 110gl(-1) CDW of E. coli. Compared to literature data, some of these cell densities are the highest values ever achieved in high cell density cultivation of microorganisms in stirred tank bioreactors. By comparing the specific power inputs as well as the k(L)a values of both systems, it is demonstrated that only the pressure is a scaleable tool for oxygen transfer enhancement in industrial stirred tank bioreactors. Furthermore, it was shown that increased carbon dioxide partial pressures did not remarkably inhibit the growth of the investigated model organisms.     

An integrated solar and artificial light system for internal illumination of photobioreactors

Exploitation of photosynthetic cells for the production of useful metabolites requires efficient photobioreactors. Many laboratory scale photobioreactors have been reported but most of them are extremely difficult to scale up. Furthermore, the use of open ponds and outdoor tubular photobioreactors is limited by the requirement for large spaces and the difficulty in maintaining sterile conditions. In view of this, we have designed and constructed an internally illuminated stirred tank photobioreactor. The photobioreactor is simple, heat sterilizable and mechanically agitated like the conventional stirred tank bioreactors. Furthermore, it can easily be scaled up while maintaining the light supply coefficient and thus the productivity constant. A device was installed for collecting solar light and distributing it inside the reactor through optical fibers. It was equipped with a light tracking sensor so that the lenses rotate with the position of the sun. This makes it possible to use solar light for photosynthetic cell cultivation in indoor photobioreactors. As a solution to the problems of night biomass loss and low productivity on cloudy days, an artificial light source was coupled with the solar light collecting device. A light intensity sensor monitors the solar light intensity and the artificial light is automatically switched on or off, depending on the solar light intensity. In this way, continuous light supply to the reactor is achieved by using solar light during sunny period, and artificial light at night and on cloudy days.     

Production of ethanol by a stirred catalytic basket reactor with immobilized yeast cells

A stirred catalytic basket reactor with immobilized yeast cells was used for the batchwise production of ethanol. Fractional conversions up to 0.99 in 10 h were attained, depending on the agitation rates, initial glucose, and cell densities. The volumetric productivity of the reactor was considerably better than that of conventional stirred tank reactors. Productivities were strongly dependent on the stirred speed.     

Comparison of two methods for designing calorimeters using stirred tank reactors

Calorimetry is a robust method for online monitoring and controlling bioprocesses in stirred tank reactors. Up to now, reactor calorimeters have not been optimally constructed for pilot scale applications. Thus, the objective of this paper is to compare two different ways for designing reactor calorimeters and validate them. The “heat capacity” method based on the mass flow of the cooling liquid in the jacket was compared with the “heat transfer” method based on the heat transfer coefficient continuously measured in the cultivation of Escherichia coli VH33 in a 50 L stirred tank reactor. It was found that the values of the “heat transfer” method agreed very well with the calculated values from the oxygen consumption. By contrast, the curve of the “heat capacity” method deviated from that of the oxygen consumption calculated with the oxycaloric equivalent. In conclusion, the “heat transfer” method has been proven to have a higher degree of validity than the “heat capacity” method. Thus, it is a better and more robust means to measure heat generation of fermentations in stirred tank bioreactors on a pilot scale. Biotechnol. Bioeng. 2013; 110: 180–190. © 2012 Wiley Periodicals, Inc.