Bioprocess Biosyst Eng DOI 10.1007/s00449-015-1409-4
ORIGINAL PAPER
Evaluation of fluorimetric pH sensors for bioprocess monitoring at low pH Nils H. Janzen1 • Michael Schmidt1 • Christian Krause2 • Dirk Weuster-Botz1
Received: 4 December 2014 / Accepted: 28 April 2015 Ó Springer-Verlag Berlin Heidelberg 2015
Abstract Optical chemical sensors are the standard for pH monitoring in small-scale bioreactors such as microtiter plates, shaking flasks or other single-use bioreactors. The dynamic pH range of the so far commercially available fluorescent pH sensors applied in small-scale bioreactors is restricted to pH monitoring around neutral pH, although many fermentation processes are performed at pH \ 6 on industrial scale. Thus, two new prototype acidic fluorescence pH sensors immobilized in single-use stirred-tank bioreactors, one with excitation at 470 nm and emission at 550 nm (sensor 470/550) and the other with excitation at 505 nm and emission at 600 nm (sensor 505/600), were characterized with respect to dynamic ranges and operational stability in representative fermentation media. Best resolution and dynamic range was observed with pH sensor 505/600 in mineral medium (dynamic range of 3.9 \ pH \ 7.2). Applying the same pH sensors to complex medium results in a drastic reduction of resolution and dynamic ranges. Yeast extract in complex medium was found to cause background fluorescence at the sensors’ operating wavelength combinations. Optical isolation of the sensor by adding a black colored polymer layer above the sensor spot and fixing an aperture made of adhesive photoresistant foil between the fluorescence reader and the transparent bottom of the polystyrene reactors enabled full & Nils H. Janzen [email protected] & Dirk Weuster-Botz [email protected] 1
Lehrstuhl fu¨r Bioverfahrenstechnik, Technische Universita¨t Mu¨nchen, Boltzmannstr. 15, 85748 Garching, Germany
2
PreSens Precision Sensing GmbH, Josef-Engert-Strasse 11, 93053 Regensburg, Germany
re-establishment of the sensor’s characteristics. Reliability and operational stability of sensor 505/600 was shown by online pH monitoring (4.5 \ pH \ 5.8) of parallel anaerobic batch fermentations of Clostridium acetobutylicum for the production of acetone, butanol and ethanol (ABE) with offline pH measurements with a standard glass electrode as reference. Keywords Bioprocess monitoring pH microsensor Fluorescence Stirred-tank microbioreactor ABE fermentation
Introduction Advanced high-throughput bioprocess development makes use of miniaturized, parallelized and automated single-use bioreactors of different scales. Most of all micro-bioreactors on a submilliliter-scale are applied like shaken microtiter plates or mini-bioreactors on a 1–10 milliliter scale [1–3]. Individual and reliable online monitoring and control of pH and dissolved oxygen (DO) is essential for extracting relevant bioprocess information out of the huge amount of process data which can be generated in parallel screening and bioprocess development systems [4]. Usually optical chemical sensors are applied for DO and pH monitoring due to the small size of the miniaturized singleuse bioreactors. Optical chemical sensors are cheap, do not require a separate reference sensor and can easily be miniaturized [5]. Most of the commercial available optical sensors for bioprocess applications use dual lifetime referencing [6], because in a first approximation lifetime is not affected by intensity or wavelength-dependent interferences [7]: Two luminophores with different lifetimes, the reference standard and the fluorescent pH indicator, are
123
Bioprocess Biosyst Eng
entrapped in a solid immobilization matrix (sensor spot) attached to the inner bioreactor wall. Hydronium ions diffuse into the sensor spot, which is in direct contact with the fermentation medium. Both luminophores can be excited simultaneously with one light source (LED) at defined wavelengths and the emission is detected with a single photodetector (photo diode). LED and photodetector can be applied contactless from the outside of the bioreactors if the reactor material is optically transparent for the applied wavelengths. Dual lifetime referencing is used to overcome the limited long-term stability of luminophores caused by photobleaching or leaching [5]. A disadvantage of optical pH sensors is the narrow dynamic pH range of pKa ± 1.5 (Ka is the acid dissociation constant of the immobilized fluorescent pH indicator). The dynamic pH range of the so far commercially available fluorescent pH sensors applied in small-scale bioreactors is thus restricted to the pH measurement around neutral pH (e.g., [8]), because many fermentation processes are controlled within this dynamic pH range. But for the fermentative production of organic acids and solvents, acidic pH is preferred to reduce contamination risks on a large scale or to produce the organic acids instead of their salts to reduce downstream processing costs. Furthermore, the pH optimum of a significant number of yeast strains of industrial relevance such as Pichia pastoris and Saccharomyces cerevisiae is between pH 4.5 and pH 6.5 [9]. Due to the lack of appropriate sensors, the aforementioned research fields are inaccessible to disposable culture vessels relying on optical pH monitoring. Thus, the focus of this paper is on the evaluation of fluorimetric pH sensors for bioprocess monitoring at low pH (pKa around 5). The prototype pH sensors will be evaluated immobilized in single-use stirred-tank bioreactors for the high-performance cultivation of bacteria [10, 11], yeasts [12] and mycelium forming microorganisms [13, 14] as well as for enzymatic processes [15], which can be operated in a bioreaction block with a magnetic inductive drive [16, 17]. This mini-bioreactor system for parallel operation of 48 stirred-tank bioreactors is equipped with dual lifetime referencing fluorescence readers for each bioreactor with two excitation wavelengths (470/505 nm) and two photodetectors sensitive at the emission wavelengths of 550 nm and at 600 nm. Thus, two different prototype fluorescence pH sensors can be evaluated: one with excitation at 470 nm and emission at 550 nm and the other with excitation at 505 nm and emission at 600 nm. The prototype sensors will be evaluated with respect to detection limits and operational stability in a representative defined mineral salts fermentation medium and a representative complex fermentation medium. The standard batch process for the anaerobic production of acetone, butanol and ethanol (ABE) with Clostridium
123
acetobutylicum is chosen as an application example because the pH in the fermentation medium covers a typical dynamic range around pH 5, e.g., pH drops down during acetate and butyrate production and increases again in the solventogenic phase [18].
Materials and methods Prototype fluorimetric pH sensors Sterilized single-use bioreactors made of polystyrene with immobilized prototype fluorimetric pH sensors were delivered by PreSens GmbH, Regensburg, Germany. The pH sensors consist of a highly ion permeable polymer matrix (basic material for sensors—BMS, PreSens GmbH, Regensburg, Germany) directly immobilized on the inner wall of the bioreactor. The reference dye and the indicator dye are immobilized in this layer. For pH sensors 470/550, a ruthenium-based reference dye (BMS low-pH-blue-v1, PreSens GmbH, Regensburg, Germany) was used while for pH sensors 505/600 a platinum-based reference dye (BMS low-pH-green-v2, PreSens GmbH, Regensburg, Germany) was applied. Both indicator dyes are based on partially chlorinated fluorescein, the dye for the low-pH-green-v2 sensor has additional methyl groups [19]. Dynamic ranges and operational stability were measured with the baffled bioreactors filled with 10 mL aqueous fermentation medium and operated in a magnetic inductive drive (bioREACTOR 48, 2 mag AG, Munich, Germany) with gas-inducing stirrers at 2800 rpm, 37 °C, and 0.1 L min-1 sterile air gassing of the head space of each reactor to ensure sterile operation. Part of the gas is suckedin by the rotating gas-inducing stirrers into the liquid phase as function of stirrer speed. Headspace cooling was set to 2 °C to minimize evaporation. Fluorimetric readers (MCR 8*2 v5, PreSens GmbH, Regensburg, Germany) were used for excitation and readout of each of the optical sensors every 30 s. The fluorimetric readers use the phase detection method: luminescence lifetimes are detected in the frequency domain [6, 17]. Phase angle shift measurements are based on the mean average of 5 consecutive readouts covering 2.5 min of two individual sensors (reactors) if not mentioned otherwise. The parameters of a sigmoidal function were estimated by minimization of the sum of squared residuals to correlate phase angle measurements with pH. Because three different sensor (reactor) batches were calibrated simultaneously, the pH measurements are based on the mean average of six individual probes (three batches with two sensors each) if not mentioned otherwise. The pH set-points of the fermentation media under study were initially adjusted with phosphoric acid or sodium hydroxide solution. For control of pH drift, 450 lL samples
Bioprocess Biosyst Eng
were taken automatically by a liquid handler (Freedom Evo, Tecan, Crailsheim, Germany) every 12 h and distributed in microtiter plates. Reference pH of the samples was measured offline with a standard glass electrode (N 6000 A, Schott Instruments, Mainz, Germany). Characterization of the medium fluorescence Excitation and emission spectra of complex medium were determined from 400 nm till 650 nm utilizing a fluorescence spectrometer (Lumina, Thermo Scientific, Waltham, USA). Parallel batch cultivation of Clostridium acetobutylicum in stirred-tank bioreactors Precultures were obtained from spore suspensions inoculated to anaerobic clostridial growth medium in anaerobic flasks with a liquid volume of 5 mL and pasteurized at 80 °C for 10 min. After an initial growth phase of 16 h at 37 °C, the germed culture suspension was transferred to a modified mineral salt 2-(N-morpholino) ethanesulfonic acid medium (MS-MES medium, 10 % v/v). MS-MES medium was prepared in anaerobic flasks (45 mL working volume) with pH 5.3 adjusted with KOH modified from [20]. The exponentially growing microorganisms were transferred into fresh MS-MES medium in anaerobic flasks after 16 h. This procedure was repeated until sufficient cells were available for the inoculation of the stirred-tank bioreactors. Before inoculation, the optical density (OD) was measured at 600 nm. Then, the inoculation volume was adjusted to achieve an initial dry cell weight concentration of 0.15 g L-1 in the stirred-tank bioreactors filled with sterile MS-MES medium. After inoculation, the initial pH was manually adjusted with NH4OH. The bioREACTOR 48 system (2mag AG, Munich, Germany) was applied with sterilized single-use baffled bioreactors made of polystyrene with prototype fluorimetric pH sensors. To ensure anaerobic reaction conditions, the system and all necessary components were stored in an anaerobic chamber overnight before each of the parallel bioreactors was filled with 12 mL of an inoculated MSMES medium with 0.1 mL L-1 polypropylene glycol as an antifoaming agent. The inoculated parallel bioreactor system was manually transferred to the control station outside the glove box. Immediately afterwards the head spaces of each of the parallel bioreactors were rinsed with 1 L h-1 of N2 for 1 h to ensure anaerobic conditions. Afterwards, the nitrogen gas supply rate of each reactor was reduced to 0.125 L h-1, the cultivation temperature was adjusted to 37 °C, and headspace cooling was set to 2 °C to reduce evaporation. Stirrer speed was adjusted to 400 rpm.
Media Mineral salt medium A mineral salt medium derived from Korz et al. [21] was used, with following variations: no thiamine was added and a different antifoam agent (Antifoam 204, Sigma Aldrich, Steinheim, Germany) was used with concentrations of 0.1–1 % (w/v). The glucose concentration was set to 25 g L-1. Clostridial growth medium Glucose, 2.5 g L-1; KH2PO4, 0.75 g L-1; K2HPO4, 0.75 g L-1; MgSO47 H2O, 0.4 g L-1; MnSO4H2O, 0.01 g L-1; FeSO47 H2O, 0.01 g L-1; NaCl, 1 g L-1; (NH4)2SO4, 2 g L-1; yeast extract, 5 g L-1; asparagine, 2 g L-1; pH 6.6 adjusted with NH4OH. MS-MES Medium: Glucose (60 g L-1), KH2PO4 (0.55 g L-1), K2HPO4 (0.42 g L-1), MgSO47 H2O (0.22 g L-1), FeSO47 H2O (0.011 g L-1), 0.08 mg L-1 biotin and 8 mg L-1 p-aminobenzoic acid, (NH4)2SO4 (5.496 g L-1) and acetic acid (2.3 g L-1). Complex medium Sensor calibrations were carried out with a complex medium consisting of 20 g L-1 peptone (Carl Roth GmbH, Karlsruhe, Germany), 20 g L-1 yeast extract (DHW, Hamburg, Germany), 30 g L-1 glucose, 13.0 g L-1 KH2PO4, 10 g L-1 K2HPO4, 6 g L-1 H2PO4 2 H2O, 5 g L-1 NaCl, 2 g L-1 (NH4)2SO4,1 g L-1 MgSO4 7 H2O and 0.2 g L-1 NH4Cl. All medium components with the exception of glucose were mixed with pure water, pH was adjusted to 6.8 with NH4OH and the medium was autoclaved (20 min at 120 °C). Separately autoclaved glucose concentrate and 0.1 mL L-1 of separately sterilized antifoam agent (Clerol FBA 265, Cognis GmbH, Du¨sseldorf, Germany) were added aseptically after sterilization.
Results and discussion Detection limits and operational stability of prototype fluorimetric pH sensors Measured phase angles after 12 h of the two prototype pH sensors 470/550 and 505/600 (excitation/emission) are plotted as function of pH (Fig. 1). The dynamic range is estimated with the condition that the standard deviation of the phase angle measurements should not exceed a DpH of ±0.2 (n = 2). Dynamic range and highest resolution are function of the media applied (Table 1). The best
123
Bioprocess Biosyst Eng Fig. 1 Phase angles (filled diamond) of two different fluorimetric pH sensors immobilized in single-use stirred-tank bioreactors after 12 h in defined mineral salt medium and in complex medium as function of pH. Standard deviations of the estimated sigmoid functions are indicated (gray lines). Inflection points (circle) and dynamic ranges are indicated as well (gray shading). Dashed lines mark the calibration data of the manufacturer (plain buffer solutions)
Table 1 Dynamic ranges of the fluorimetric pH sensors with the corresponding phase angles
Medium
Sensor
Dynamic range pH
Phase angle
Min. (-) Mineral salt Complex
Inflection point
Max. (-)
Max. (°)
pH (-)
Slope (°)
Min. (°)
470/550
3.0
6.4
48.65
28.66
4.46
-9.2
505/600
3.9
7.2
51.51
13.82
5.51
-18.4
470/550
4.0
5.9
33.25
28.35
4.01
-6.5
505/600
5.0
7.0
42.61
14.89
6.08
-20.3
The inflection point offers the best resolution due to maximal slope of the respective curve
resolution was observed with pH sensor 505/600 in mineral medium (dynamic range 3.9 \ pH \ 7.2). But applying the same pH sensor to the complex medium results in a drastic reduction of resolution and dynamic range (5.0 \ pH \ 7.0). The same behavior can be observed with the less sensitive pH sensor 470/550 (Fig. 1). Both prototype pH sensors show significant deviations compared to the calibration data in plain buffer solutions (provided by the manufacturer). Within the observation time of 72 h, the sensors drifted only significantly between the initial phase angle measurements and the measurements after 12 h (Fig. 2). The initial drift was less distinct with the mineral medium.
123
During an initial period of up to 12 h after first wetting with media, the sensor has to swell and to equilibrate with media components. In case of mineral medium, this process affects the signal of the sensors less than with complex media. The supposed reason is an accumulation of fluorescent media components at the sensor surface. They create a background fluorescence which results in a decrease of the upper asymptote of the sigmoidal calibration curve (Fig. 1) and therefore a decrease in resolution and dynamic range. The fluorescence characteristic of the complex medium was studied due to the compromised performance of both fluorimetric pH sensors in complex medium (Fig. 3). Excitation of the complex medium is highest at 450 nm and
Bioprocess Biosyst Eng Fig. 2 Fitted sigmoidal functions of two different fluorimetric pH sensors immobilized in single-use stirred-tank bioreactors filled with a defined mineral salt medium and a complex medium as function of pH. Sigmoidal functions were identified every 12 h within a period of 72 h. The initial estimate (t = 0 h) is indicated in dotted lines and the final estimate in black (t = 72 h)
Fig. 3 Excitation spectrum of complex medium (pH 6.8) with indication of the excitation wavelengths of the fluorimetric sensors (a) and emission intensities of complex medium after excitation with 470 nm (gray) and 505 nm (black) with indication of the readout wavelengths of the fluorimetric sensors at 550 and 600 nm
decreases with increasing wavelengths. Excitation intensity at the excitation wavelength of 505 nm is 12.7 % of the excitation intensity at the excitation wavelength of 470 nm. The emission spectrum of the complex medium after excitation with 470 nm shows a maximum nearby the detection wavelength of 550 nm of the corresponding pH sensor, whereas the emission spectrum after excitation with 505 nm is very much reduced at the corresponding detection wavelength of 600 nm (factor 13 compared to 470/550). Only negligible luminescence activities were
observed without the complex components yeast extract and peptone (data not shown). Some of the fluorimetric pH sensors 505/600 were equipped with an additional black colored hydrogel layer on top of the fluorescence sensor material to reduce the compromising background fluorescence. Comparative phase angle measurements were performed as function of pH with single-use bioreactors filled with the complex medium. Additionally, an aperture made of adhesive photoresistant foil was applied beneath the bottom of the
123
Bioprocess Biosyst Eng
Fig. 4 Measured phase angles (filled diamond) of fluorimetric pH sensor 505/600 (dashed lines) in mineral salt medium (left) and in complex medium (right) together with its optically isolated variant with (continuous lines) and without (dashed lines) application of an aperture of adhesive photoresistant foil between fluorescence reader
and single-use bioreactor after 12 h in complex medium. Sigmoid functions were fitted to the measured values. For each sigmoid function, the standard deviation (gray lines) is shown. Dynamic ranges are indicated as well (gray shading)
single-use bioreactors to further shield the fluorescence readers below from the medium’s background fluorescence emitted through the transparent bottom of the polystyrene reactors. Calibration curves were measured with all configurations (Fig. 4). The application of an additional black hydrogel layer on top of the sensors in combination with the aperture foil resulted in full re-establishment of the sensor’s characteristics like in the mineral salt medium.
acidification was observed in one of the parallel fermentations (Fig. 5a). In the course of the fermentation, the standard deviation of the optical sensors’ measurements increases due to the individual pH variance of the multifold batch processes. Offline reference measurements with the glass electrode of individual reactors are within the standard deviations of the online measured pH.
Bioprocess application of the pH sensors 505/600 (ABE fermentation)
Conclusions
Sensor 505/600 without any optical isolation was selected to monitor dynamic pH changes in parallel anaerobic batch processes for ABE production with Clostridium acetobutylicum in a mineral medium. The identified sigmoidal function of the MS-MES medium was used as calibration function for online pH recording in 24 parallel stirred-tank bioreactors. Online pH data are plotted with a step time of 6 min for better visualization of the standard deviations (Fig. 5). Two different parallel fermentations are shown with varying initial pH. The first parallel 24-fold fermentations were adjusted to an initial pH of 5.5 according to [18], while a second parallel batch was adjusted to an initial pH of 5.8. Reference pH measurements were performed with a standard glass electrode by draining one of the parallel reactors for each individual measurement. Thus, the number of parallel reactors was reduced from initial 24 to 3 single-use bioreactors in the end. The pH drops down from initially pH 5.4 (pH 5.7) to pH 4.5 in the acidogenic phase of the fermentation processes with Clostridium acetobutylicum and increases afterwards to pH 4.8 (pH 4.9) in the solventogenic phase, before a second
A new fluorescence pH sensor with excitation at 505 nm and emission at 600 nm (pH sensor 505/600) is now available for online pH measurement between pH 4.0 and pH 7.0. The operational stability of pH sensor 505/600 was shown with representative fermentation media and by online pH monitoring (4.5 \ pH \ 5.8) of the anaerobic batch production of acetone, butanol and ethanol (ABE) with Clostridium acetobutylicum. By providing a cost-efficient and flexible way of pH monitoring at acidic pH, the new sensors will enable pH-controlled parallel screening studies of industrially relevant yeast strains and other acidophilic microorganisms for high-throughput screening in disposable vessels. The small size of the fluorimetric sensors and their non-invasive readout capabilities makes them easily applicable to other single-use bioreactors such as microwell plates, shaking flasks or other miniaturized bioreactors. Medium fluorescence was shown to be a major problem with fluorimetric pH sensors. Biomolecules like oligopeptides as well as cell debris in yeast extract were found to cause background fluorescence at the sensors’ operating wavelength combinations of 470/550 and 505/600 nm. The
123
Bioprocess Biosyst Eng Fig. 5 Parallel online pH measurements with fluorimetric pH sensors 505/600 (circle) and individual offline pH measurements with a standard glass electrode (filled diamond) of initially 24 parallel ABE fermentations with Clostridium acetobutylicum in single-use stirred-tank bioreactors and corresponding OD600 (filled delta). Two different parallel batch fermentations are shown (a, b)
massive interaction with complex components of the 470/550 nm wavelength combination is less suited for pH measurement in complex media than the less fluorescence causing 505/600 nm wavelength combination. As the sensors’ operating principle, dual lifetime referencing (DLR) relies on both the sensor material’s phosphorescence and fluorescence signal, the observed background fluorescence can cause the diminishing sensor performance. This is one of the few sources of interference that cannot be compensated by the otherwise very robust method of dual lifetime as already reported by Klimant et al. [6]. While most adverse effects are referenced out by DLR, the intrinsic background fluorescence is an exception. This was to be expected since the decay profile of background fluorescence cannot be distinguished from that of the fluorescence indicator. Optical isolation by covering the fluorimetric pH sensor with a black colored polymer layer to reduce background fluorescence of the medium and application of an aperture plate to shield the fluorescence readers from the medium’s background fluorescence emitted through the transparent bottom of the polystyrene reactors enables full re-establishment of the sensor’s characteristics like in a nonfluorescent medium. The positive effects of an optical isolation are of great value because not only oligopeptides and
cell debris do interfere with this optical measurement, but also living cells. Since most optical sensor applications use low cell densities in microtiter plates and shaking flask, the interference of the cells is negligible and therefore has not been important so far. But with high-performance stirredtank bioreactor systems like the bioREACTOR 48, high cell density fermentations are possible. Without optical isolation, fluorimetric pH measurements have been limited so far to maximum dry cell mass concentrations of *28 g L-1 E.coli [11]. Acknowledgments This work was funded by the Bavarian Ministry of Economic Affairs and Media, Energy and Technology and the companies PreSens Precision Sensing GmbH, Regensburg, Germany and 2mag AG, Munich, Germany. Assistance from Michael Geitner and software support by Dirk Hebel (Technische Universita¨t Mu¨nchen, Garching, Germany) is gratefully acknowledged. The authors also gratefully acknowledge the support of Nils H. Janzen and Michael Schmidt by the TUM Graduate School at the Technische Universita¨t Mu¨nchen, Munich, Germany.
References 1. Huber R, Ritter D, Hering T, Hillmer A-K, Kensy F, Mu¨ller C, Wang L, Bu¨chs J (2009) Robo-Lector—a novel platform for automated high-throughput cultivations in microtiter plates with
123
Bioprocess Biosyst Eng
2.
3.
4. 5. 6.
7.
8.
9. 10.
11.
high information content. Microb Cell Fact 8:42. doi:10.1186/ 1475-2859-8-42 Hortsch R, Weuster-Botz D (2010) Milliliter-scale stirred tank reactors for the cultivation of microorganisms. In Laskin AI, Sariaslani S and Gadd GM (Eds): Advances in Applied Microbiology, Vol. 73, Elsevier BV, Amsterdam, pp 61–82 Kirk T, Szita N (2012) Oxygen transfer characteristics of miniaturized bioreactor systems. Biotechnol Bioeng 110:1005–1019 Harms P, Kostov Y, Rao G (2002) Bioprocess monitoring. Curr Op Biotechnol 13:124–127 Wencel D, Abel T, McDonagh C (2014) optical chemical pH sensors. Anal Chem 86:15–29 Klimant I, Huber C, Liebsch G, Wolfbeis OS (2001): Dual lifetime referencing (DLR)—a new scheme for converting fluorescence intensity into a frequency-domain or time-domain information. In: Valeur, Bernard und Brochon, Jean-Claude, (eds.) New trends in fluorescence spectroscopy: applications to chemical and life sciences. Springer Series on Fluorescence, 1 (13). Springer, Berlin, 257-274 Liebsch G, Klimant I, Krause C, Wolfbeis OS (2001) Fluorescent imaging of pH with optical sensors using time domain dual lifetime referencing. Anal Chem 73:4354–4363 Arain A, John GT, Krause C, Gerlach J, Wolfbeis OS, Klimant I (2006) Characterization of microtiterplates with integrated optical sensors for oxygen and pH, and their applications to enzyme activity screening, respirometry, and toxicological assays. Sens Actuat B 113:639–648 Walker GM (2000) Yeast physiology and biotechnology. Wiley, West Sussex, pp 144–149 Knorr B, Schlieker H, Hohmann H-P, Weuster-Botz D (2007) Scale-down and parallel operation of the riboflavin production process with Bacillus subtilis. Biochem Eng J 33:263–274 Faust G, Janzen N, Bendig C, Ro¨mer L, Kaufmann K, WeusterBotz D (2014) Feeding strategies enhance high cell density
123
12.
13.
14.
15.
16.
17.
18.
19. 20.
21.
cultivation and protein expression in milliliter-scale bioreactors. Biotechnol J 9:1293–1303 Kusterer A, Krause C, Kaufmann K, Arnold M, Weuster-Botz D (2008) Fully automated single-use stirred-tank bioreactors for parallel microbial cultivations. Bioproc Biosys Eng 31:207–215 Hortsch R, Krispin H, Weuster-Botz D (2011) Process performance of parallel bioreactors for batch cultivation of Streptomyces tendae. Bioproc Biosys Eng 34:297–304 Bendig C, Weuster-Botz D (2013) Reaction engineering analysis of cellulase production with Trichoderma reesei RUT-C30 with intermittent substrate supply. Bioproc Biosys Eng 36:893–900 Riedlberger P, Bru¨ning S, Weuster-Botz D (2013) Characterization of stirrers for screening studies of enzymatic biomass hydrolyses on a milliliter-scale. Bioproc Biosys Eng 36:927–935 Puskeiler R, Kaufmann K, Weuster-Botz D (2005) Development, parallelization and automation of a gas-inducing milliliter-scale bioreactor for high-throughput bioprocess design (HTBD). Biotechnol Bioeng 89:512–523 Weuster-Botz D, Puskeiler R, Kusterer A, Kaufmann K, John G, Arnold M (2005) Methods and milliliter scale devices for highthroughput bioprocess design. Bioproc Biosys Eng 28:109–119 Schmidt M, Weuster-Botz D (2012) Reaction engineering studies of acetone-butanol-ethanol fermentation with Clostridium acetobutylicum. Biotechnol J 7:656–661 Klimant I (2003) Patent US 6,602,716 B1 Monot F, Martin JR, Petitdemange H, Gay R (1982) Acetone and butanol production by Clostridium acetobutylicum in a synthetic medium. Appl Environ Microbiol 44:1318–1324 Korz DJ, Rinas U, Hellmuth K, Sanders EA, Deckwer WD (1995) Simple fed-batch technique for high cell density cultivation of Escherichia coli. J Biotechnol 39:59–65
ORIGINAL PAPER
Evaluation of fluorimetric pH sensors for bioprocess monitoring at low pH Nils H. Janzen1 • Michael Schmidt1 • Christian Krause2 • Dirk Weuster-Botz1
Received: 4 December 2014 / Accepted: 28 April 2015 Ó Springer-Verlag Berlin Heidelberg 2015
Abstract Optical chemical sensors are the standard for pH monitoring in small-scale bioreactors such as microtiter plates, shaking flasks or other single-use bioreactors. The dynamic pH range of the so far commercially available fluorescent pH sensors applied in small-scale bioreactors is restricted to pH monitoring around neutral pH, although many fermentation processes are performed at pH \ 6 on industrial scale. Thus, two new prototype acidic fluorescence pH sensors immobilized in single-use stirred-tank bioreactors, one with excitation at 470 nm and emission at 550 nm (sensor 470/550) and the other with excitation at 505 nm and emission at 600 nm (sensor 505/600), were characterized with respect to dynamic ranges and operational stability in representative fermentation media. Best resolution and dynamic range was observed with pH sensor 505/600 in mineral medium (dynamic range of 3.9 \ pH \ 7.2). Applying the same pH sensors to complex medium results in a drastic reduction of resolution and dynamic ranges. Yeast extract in complex medium was found to cause background fluorescence at the sensors’ operating wavelength combinations. Optical isolation of the sensor by adding a black colored polymer layer above the sensor spot and fixing an aperture made of adhesive photoresistant foil between the fluorescence reader and the transparent bottom of the polystyrene reactors enabled full & Nils H. Janzen [email protected] & Dirk Weuster-Botz [email protected] 1
Lehrstuhl fu¨r Bioverfahrenstechnik, Technische Universita¨t Mu¨nchen, Boltzmannstr. 15, 85748 Garching, Germany
2
PreSens Precision Sensing GmbH, Josef-Engert-Strasse 11, 93053 Regensburg, Germany
re-establishment of the sensor’s characteristics. Reliability and operational stability of sensor 505/600 was shown by online pH monitoring (4.5 \ pH \ 5.8) of parallel anaerobic batch fermentations of Clostridium acetobutylicum for the production of acetone, butanol and ethanol (ABE) with offline pH measurements with a standard glass electrode as reference. Keywords Bioprocess monitoring pH microsensor Fluorescence Stirred-tank microbioreactor ABE fermentation
Introduction Advanced high-throughput bioprocess development makes use of miniaturized, parallelized and automated single-use bioreactors of different scales. Most of all micro-bioreactors on a submilliliter-scale are applied like shaken microtiter plates or mini-bioreactors on a 1–10 milliliter scale [1–3]. Individual and reliable online monitoring and control of pH and dissolved oxygen (DO) is essential for extracting relevant bioprocess information out of the huge amount of process data which can be generated in parallel screening and bioprocess development systems [4]. Usually optical chemical sensors are applied for DO and pH monitoring due to the small size of the miniaturized singleuse bioreactors. Optical chemical sensors are cheap, do not require a separate reference sensor and can easily be miniaturized [5]. Most of the commercial available optical sensors for bioprocess applications use dual lifetime referencing [6], because in a first approximation lifetime is not affected by intensity or wavelength-dependent interferences [7]: Two luminophores with different lifetimes, the reference standard and the fluorescent pH indicator, are
123
Bioprocess Biosyst Eng
entrapped in a solid immobilization matrix (sensor spot) attached to the inner bioreactor wall. Hydronium ions diffuse into the sensor spot, which is in direct contact with the fermentation medium. Both luminophores can be excited simultaneously with one light source (LED) at defined wavelengths and the emission is detected with a single photodetector (photo diode). LED and photodetector can be applied contactless from the outside of the bioreactors if the reactor material is optically transparent for the applied wavelengths. Dual lifetime referencing is used to overcome the limited long-term stability of luminophores caused by photobleaching or leaching [5]. A disadvantage of optical pH sensors is the narrow dynamic pH range of pKa ± 1.5 (Ka is the acid dissociation constant of the immobilized fluorescent pH indicator). The dynamic pH range of the so far commercially available fluorescent pH sensors applied in small-scale bioreactors is thus restricted to the pH measurement around neutral pH (e.g., [8]), because many fermentation processes are controlled within this dynamic pH range. But for the fermentative production of organic acids and solvents, acidic pH is preferred to reduce contamination risks on a large scale or to produce the organic acids instead of their salts to reduce downstream processing costs. Furthermore, the pH optimum of a significant number of yeast strains of industrial relevance such as Pichia pastoris and Saccharomyces cerevisiae is between pH 4.5 and pH 6.5 [9]. Due to the lack of appropriate sensors, the aforementioned research fields are inaccessible to disposable culture vessels relying on optical pH monitoring. Thus, the focus of this paper is on the evaluation of fluorimetric pH sensors for bioprocess monitoring at low pH (pKa around 5). The prototype pH sensors will be evaluated immobilized in single-use stirred-tank bioreactors for the high-performance cultivation of bacteria [10, 11], yeasts [12] and mycelium forming microorganisms [13, 14] as well as for enzymatic processes [15], which can be operated in a bioreaction block with a magnetic inductive drive [16, 17]. This mini-bioreactor system for parallel operation of 48 stirred-tank bioreactors is equipped with dual lifetime referencing fluorescence readers for each bioreactor with two excitation wavelengths (470/505 nm) and two photodetectors sensitive at the emission wavelengths of 550 nm and at 600 nm. Thus, two different prototype fluorescence pH sensors can be evaluated: one with excitation at 470 nm and emission at 550 nm and the other with excitation at 505 nm and emission at 600 nm. The prototype sensors will be evaluated with respect to detection limits and operational stability in a representative defined mineral salts fermentation medium and a representative complex fermentation medium. The standard batch process for the anaerobic production of acetone, butanol and ethanol (ABE) with Clostridium
123
acetobutylicum is chosen as an application example because the pH in the fermentation medium covers a typical dynamic range around pH 5, e.g., pH drops down during acetate and butyrate production and increases again in the solventogenic phase [18].
Materials and methods Prototype fluorimetric pH sensors Sterilized single-use bioreactors made of polystyrene with immobilized prototype fluorimetric pH sensors were delivered by PreSens GmbH, Regensburg, Germany. The pH sensors consist of a highly ion permeable polymer matrix (basic material for sensors—BMS, PreSens GmbH, Regensburg, Germany) directly immobilized on the inner wall of the bioreactor. The reference dye and the indicator dye are immobilized in this layer. For pH sensors 470/550, a ruthenium-based reference dye (BMS low-pH-blue-v1, PreSens GmbH, Regensburg, Germany) was used while for pH sensors 505/600 a platinum-based reference dye (BMS low-pH-green-v2, PreSens GmbH, Regensburg, Germany) was applied. Both indicator dyes are based on partially chlorinated fluorescein, the dye for the low-pH-green-v2 sensor has additional methyl groups [19]. Dynamic ranges and operational stability were measured with the baffled bioreactors filled with 10 mL aqueous fermentation medium and operated in a magnetic inductive drive (bioREACTOR 48, 2 mag AG, Munich, Germany) with gas-inducing stirrers at 2800 rpm, 37 °C, and 0.1 L min-1 sterile air gassing of the head space of each reactor to ensure sterile operation. Part of the gas is suckedin by the rotating gas-inducing stirrers into the liquid phase as function of stirrer speed. Headspace cooling was set to 2 °C to minimize evaporation. Fluorimetric readers (MCR 8*2 v5, PreSens GmbH, Regensburg, Germany) were used for excitation and readout of each of the optical sensors every 30 s. The fluorimetric readers use the phase detection method: luminescence lifetimes are detected in the frequency domain [6, 17]. Phase angle shift measurements are based on the mean average of 5 consecutive readouts covering 2.5 min of two individual sensors (reactors) if not mentioned otherwise. The parameters of a sigmoidal function were estimated by minimization of the sum of squared residuals to correlate phase angle measurements with pH. Because three different sensor (reactor) batches were calibrated simultaneously, the pH measurements are based on the mean average of six individual probes (three batches with two sensors each) if not mentioned otherwise. The pH set-points of the fermentation media under study were initially adjusted with phosphoric acid or sodium hydroxide solution. For control of pH drift, 450 lL samples
Bioprocess Biosyst Eng
were taken automatically by a liquid handler (Freedom Evo, Tecan, Crailsheim, Germany) every 12 h and distributed in microtiter plates. Reference pH of the samples was measured offline with a standard glass electrode (N 6000 A, Schott Instruments, Mainz, Germany). Characterization of the medium fluorescence Excitation and emission spectra of complex medium were determined from 400 nm till 650 nm utilizing a fluorescence spectrometer (Lumina, Thermo Scientific, Waltham, USA). Parallel batch cultivation of Clostridium acetobutylicum in stirred-tank bioreactors Precultures were obtained from spore suspensions inoculated to anaerobic clostridial growth medium in anaerobic flasks with a liquid volume of 5 mL and pasteurized at 80 °C for 10 min. After an initial growth phase of 16 h at 37 °C, the germed culture suspension was transferred to a modified mineral salt 2-(N-morpholino) ethanesulfonic acid medium (MS-MES medium, 10 % v/v). MS-MES medium was prepared in anaerobic flasks (45 mL working volume) with pH 5.3 adjusted with KOH modified from [20]. The exponentially growing microorganisms were transferred into fresh MS-MES medium in anaerobic flasks after 16 h. This procedure was repeated until sufficient cells were available for the inoculation of the stirred-tank bioreactors. Before inoculation, the optical density (OD) was measured at 600 nm. Then, the inoculation volume was adjusted to achieve an initial dry cell weight concentration of 0.15 g L-1 in the stirred-tank bioreactors filled with sterile MS-MES medium. After inoculation, the initial pH was manually adjusted with NH4OH. The bioREACTOR 48 system (2mag AG, Munich, Germany) was applied with sterilized single-use baffled bioreactors made of polystyrene with prototype fluorimetric pH sensors. To ensure anaerobic reaction conditions, the system and all necessary components were stored in an anaerobic chamber overnight before each of the parallel bioreactors was filled with 12 mL of an inoculated MSMES medium with 0.1 mL L-1 polypropylene glycol as an antifoaming agent. The inoculated parallel bioreactor system was manually transferred to the control station outside the glove box. Immediately afterwards the head spaces of each of the parallel bioreactors were rinsed with 1 L h-1 of N2 for 1 h to ensure anaerobic conditions. Afterwards, the nitrogen gas supply rate of each reactor was reduced to 0.125 L h-1, the cultivation temperature was adjusted to 37 °C, and headspace cooling was set to 2 °C to reduce evaporation. Stirrer speed was adjusted to 400 rpm.
Media Mineral salt medium A mineral salt medium derived from Korz et al. [21] was used, with following variations: no thiamine was added and a different antifoam agent (Antifoam 204, Sigma Aldrich, Steinheim, Germany) was used with concentrations of 0.1–1 % (w/v). The glucose concentration was set to 25 g L-1. Clostridial growth medium Glucose, 2.5 g L-1; KH2PO4, 0.75 g L-1; K2HPO4, 0.75 g L-1; MgSO47 H2O, 0.4 g L-1; MnSO4H2O, 0.01 g L-1; FeSO47 H2O, 0.01 g L-1; NaCl, 1 g L-1; (NH4)2SO4, 2 g L-1; yeast extract, 5 g L-1; asparagine, 2 g L-1; pH 6.6 adjusted with NH4OH. MS-MES Medium: Glucose (60 g L-1), KH2PO4 (0.55 g L-1), K2HPO4 (0.42 g L-1), MgSO47 H2O (0.22 g L-1), FeSO47 H2O (0.011 g L-1), 0.08 mg L-1 biotin and 8 mg L-1 p-aminobenzoic acid, (NH4)2SO4 (5.496 g L-1) and acetic acid (2.3 g L-1). Complex medium Sensor calibrations were carried out with a complex medium consisting of 20 g L-1 peptone (Carl Roth GmbH, Karlsruhe, Germany), 20 g L-1 yeast extract (DHW, Hamburg, Germany), 30 g L-1 glucose, 13.0 g L-1 KH2PO4, 10 g L-1 K2HPO4, 6 g L-1 H2PO4 2 H2O, 5 g L-1 NaCl, 2 g L-1 (NH4)2SO4,1 g L-1 MgSO4 7 H2O and 0.2 g L-1 NH4Cl. All medium components with the exception of glucose were mixed with pure water, pH was adjusted to 6.8 with NH4OH and the medium was autoclaved (20 min at 120 °C). Separately autoclaved glucose concentrate and 0.1 mL L-1 of separately sterilized antifoam agent (Clerol FBA 265, Cognis GmbH, Du¨sseldorf, Germany) were added aseptically after sterilization.
Results and discussion Detection limits and operational stability of prototype fluorimetric pH sensors Measured phase angles after 12 h of the two prototype pH sensors 470/550 and 505/600 (excitation/emission) are plotted as function of pH (Fig. 1). The dynamic range is estimated with the condition that the standard deviation of the phase angle measurements should not exceed a DpH of ±0.2 (n = 2). Dynamic range and highest resolution are function of the media applied (Table 1). The best
123
Bioprocess Biosyst Eng Fig. 1 Phase angles (filled diamond) of two different fluorimetric pH sensors immobilized in single-use stirred-tank bioreactors after 12 h in defined mineral salt medium and in complex medium as function of pH. Standard deviations of the estimated sigmoid functions are indicated (gray lines). Inflection points (circle) and dynamic ranges are indicated as well (gray shading). Dashed lines mark the calibration data of the manufacturer (plain buffer solutions)
Table 1 Dynamic ranges of the fluorimetric pH sensors with the corresponding phase angles
Medium
Sensor
Dynamic range pH
Phase angle
Min. (-) Mineral salt Complex
Inflection point
Max. (-)
Max. (°)
pH (-)
Slope (°)
Min. (°)
470/550
3.0
6.4
48.65
28.66
4.46
-9.2
505/600
3.9
7.2
51.51
13.82
5.51
-18.4
470/550
4.0
5.9
33.25
28.35
4.01
-6.5
505/600
5.0
7.0
42.61
14.89
6.08
-20.3
The inflection point offers the best resolution due to maximal slope of the respective curve
resolution was observed with pH sensor 505/600 in mineral medium (dynamic range 3.9 \ pH \ 7.2). But applying the same pH sensor to the complex medium results in a drastic reduction of resolution and dynamic range (5.0 \ pH \ 7.0). The same behavior can be observed with the less sensitive pH sensor 470/550 (Fig. 1). Both prototype pH sensors show significant deviations compared to the calibration data in plain buffer solutions (provided by the manufacturer). Within the observation time of 72 h, the sensors drifted only significantly between the initial phase angle measurements and the measurements after 12 h (Fig. 2). The initial drift was less distinct with the mineral medium.
123
During an initial period of up to 12 h after first wetting with media, the sensor has to swell and to equilibrate with media components. In case of mineral medium, this process affects the signal of the sensors less than with complex media. The supposed reason is an accumulation of fluorescent media components at the sensor surface. They create a background fluorescence which results in a decrease of the upper asymptote of the sigmoidal calibration curve (Fig. 1) and therefore a decrease in resolution and dynamic range. The fluorescence characteristic of the complex medium was studied due to the compromised performance of both fluorimetric pH sensors in complex medium (Fig. 3). Excitation of the complex medium is highest at 450 nm and
Bioprocess Biosyst Eng Fig. 2 Fitted sigmoidal functions of two different fluorimetric pH sensors immobilized in single-use stirred-tank bioreactors filled with a defined mineral salt medium and a complex medium as function of pH. Sigmoidal functions were identified every 12 h within a period of 72 h. The initial estimate (t = 0 h) is indicated in dotted lines and the final estimate in black (t = 72 h)
Fig. 3 Excitation spectrum of complex medium (pH 6.8) with indication of the excitation wavelengths of the fluorimetric sensors (a) and emission intensities of complex medium after excitation with 470 nm (gray) and 505 nm (black) with indication of the readout wavelengths of the fluorimetric sensors at 550 and 600 nm
decreases with increasing wavelengths. Excitation intensity at the excitation wavelength of 505 nm is 12.7 % of the excitation intensity at the excitation wavelength of 470 nm. The emission spectrum of the complex medium after excitation with 470 nm shows a maximum nearby the detection wavelength of 550 nm of the corresponding pH sensor, whereas the emission spectrum after excitation with 505 nm is very much reduced at the corresponding detection wavelength of 600 nm (factor 13 compared to 470/550). Only negligible luminescence activities were
observed without the complex components yeast extract and peptone (data not shown). Some of the fluorimetric pH sensors 505/600 were equipped with an additional black colored hydrogel layer on top of the fluorescence sensor material to reduce the compromising background fluorescence. Comparative phase angle measurements were performed as function of pH with single-use bioreactors filled with the complex medium. Additionally, an aperture made of adhesive photoresistant foil was applied beneath the bottom of the
123
Bioprocess Biosyst Eng
Fig. 4 Measured phase angles (filled diamond) of fluorimetric pH sensor 505/600 (dashed lines) in mineral salt medium (left) and in complex medium (right) together with its optically isolated variant with (continuous lines) and without (dashed lines) application of an aperture of adhesive photoresistant foil between fluorescence reader
and single-use bioreactor after 12 h in complex medium. Sigmoid functions were fitted to the measured values. For each sigmoid function, the standard deviation (gray lines) is shown. Dynamic ranges are indicated as well (gray shading)
single-use bioreactors to further shield the fluorescence readers below from the medium’s background fluorescence emitted through the transparent bottom of the polystyrene reactors. Calibration curves were measured with all configurations (Fig. 4). The application of an additional black hydrogel layer on top of the sensors in combination with the aperture foil resulted in full re-establishment of the sensor’s characteristics like in the mineral salt medium.
acidification was observed in one of the parallel fermentations (Fig. 5a). In the course of the fermentation, the standard deviation of the optical sensors’ measurements increases due to the individual pH variance of the multifold batch processes. Offline reference measurements with the glass electrode of individual reactors are within the standard deviations of the online measured pH.
Bioprocess application of the pH sensors 505/600 (ABE fermentation)
Conclusions
Sensor 505/600 without any optical isolation was selected to monitor dynamic pH changes in parallel anaerobic batch processes for ABE production with Clostridium acetobutylicum in a mineral medium. The identified sigmoidal function of the MS-MES medium was used as calibration function for online pH recording in 24 parallel stirred-tank bioreactors. Online pH data are plotted with a step time of 6 min for better visualization of the standard deviations (Fig. 5). Two different parallel fermentations are shown with varying initial pH. The first parallel 24-fold fermentations were adjusted to an initial pH of 5.5 according to [18], while a second parallel batch was adjusted to an initial pH of 5.8. Reference pH measurements were performed with a standard glass electrode by draining one of the parallel reactors for each individual measurement. Thus, the number of parallel reactors was reduced from initial 24 to 3 single-use bioreactors in the end. The pH drops down from initially pH 5.4 (pH 5.7) to pH 4.5 in the acidogenic phase of the fermentation processes with Clostridium acetobutylicum and increases afterwards to pH 4.8 (pH 4.9) in the solventogenic phase, before a second
A new fluorescence pH sensor with excitation at 505 nm and emission at 600 nm (pH sensor 505/600) is now available for online pH measurement between pH 4.0 and pH 7.0. The operational stability of pH sensor 505/600 was shown with representative fermentation media and by online pH monitoring (4.5 \ pH \ 5.8) of the anaerobic batch production of acetone, butanol and ethanol (ABE) with Clostridium acetobutylicum. By providing a cost-efficient and flexible way of pH monitoring at acidic pH, the new sensors will enable pH-controlled parallel screening studies of industrially relevant yeast strains and other acidophilic microorganisms for high-throughput screening in disposable vessels. The small size of the fluorimetric sensors and their non-invasive readout capabilities makes them easily applicable to other single-use bioreactors such as microwell plates, shaking flasks or other miniaturized bioreactors. Medium fluorescence was shown to be a major problem with fluorimetric pH sensors. Biomolecules like oligopeptides as well as cell debris in yeast extract were found to cause background fluorescence at the sensors’ operating wavelength combinations of 470/550 and 505/600 nm. The
123
Bioprocess Biosyst Eng Fig. 5 Parallel online pH measurements with fluorimetric pH sensors 505/600 (circle) and individual offline pH measurements with a standard glass electrode (filled diamond) of initially 24 parallel ABE fermentations with Clostridium acetobutylicum in single-use stirred-tank bioreactors and corresponding OD600 (filled delta). Two different parallel batch fermentations are shown (a, b)
massive interaction with complex components of the 470/550 nm wavelength combination is less suited for pH measurement in complex media than the less fluorescence causing 505/600 nm wavelength combination. As the sensors’ operating principle, dual lifetime referencing (DLR) relies on both the sensor material’s phosphorescence and fluorescence signal, the observed background fluorescence can cause the diminishing sensor performance. This is one of the few sources of interference that cannot be compensated by the otherwise very robust method of dual lifetime as already reported by Klimant et al. [6]. While most adverse effects are referenced out by DLR, the intrinsic background fluorescence is an exception. This was to be expected since the decay profile of background fluorescence cannot be distinguished from that of the fluorescence indicator. Optical isolation by covering the fluorimetric pH sensor with a black colored polymer layer to reduce background fluorescence of the medium and application of an aperture plate to shield the fluorescence readers from the medium’s background fluorescence emitted through the transparent bottom of the polystyrene reactors enables full re-establishment of the sensor’s characteristics like in a nonfluorescent medium. The positive effects of an optical isolation are of great value because not only oligopeptides and
cell debris do interfere with this optical measurement, but also living cells. Since most optical sensor applications use low cell densities in microtiter plates and shaking flask, the interference of the cells is negligible and therefore has not been important so far. But with high-performance stirredtank bioreactor systems like the bioREACTOR 48, high cell density fermentations are possible. Without optical isolation, fluorimetric pH measurements have been limited so far to maximum dry cell mass concentrations of *28 g L-1 E.coli [11]. Acknowledgments This work was funded by the Bavarian Ministry of Economic Affairs and Media, Energy and Technology and the companies PreSens Precision Sensing GmbH, Regensburg, Germany and 2mag AG, Munich, Germany. Assistance from Michael Geitner and software support by Dirk Hebel (Technische Universita¨t Mu¨nchen, Garching, Germany) is gratefully acknowledged. The authors also gratefully acknowledge the support of Nils H. Janzen and Michael Schmidt by the TUM Graduate School at the Technische Universita¨t Mu¨nchen, Munich, Germany.
References 1. Huber R, Ritter D, Hering T, Hillmer A-K, Kensy F, Mu¨ller C, Wang L, Bu¨chs J (2009) Robo-Lector—a novel platform for automated high-throughput cultivations in microtiter plates with
123
Bioprocess Biosyst Eng
2.
3.
4. 5. 6.
7.
8.
9. 10.
11.
high information content. Microb Cell Fact 8:42. doi:10.1186/ 1475-2859-8-42 Hortsch R, Weuster-Botz D (2010) Milliliter-scale stirred tank reactors for the cultivation of microorganisms. In Laskin AI, Sariaslani S and Gadd GM (Eds): Advances in Applied Microbiology, Vol. 73, Elsevier BV, Amsterdam, pp 61–82 Kirk T, Szita N (2012) Oxygen transfer characteristics of miniaturized bioreactor systems. Biotechnol Bioeng 110:1005–1019 Harms P, Kostov Y, Rao G (2002) Bioprocess monitoring. Curr Op Biotechnol 13:124–127 Wencel D, Abel T, McDonagh C (2014) optical chemical pH sensors. Anal Chem 86:15–29 Klimant I, Huber C, Liebsch G, Wolfbeis OS (2001): Dual lifetime referencing (DLR)—a new scheme for converting fluorescence intensity into a frequency-domain or time-domain information. In: Valeur, Bernard und Brochon, Jean-Claude, (eds.) New trends in fluorescence spectroscopy: applications to chemical and life sciences. Springer Series on Fluorescence, 1 (13). Springer, Berlin, 257-274 Liebsch G, Klimant I, Krause C, Wolfbeis OS (2001) Fluorescent imaging of pH with optical sensors using time domain dual lifetime referencing. Anal Chem 73:4354–4363 Arain A, John GT, Krause C, Gerlach J, Wolfbeis OS, Klimant I (2006) Characterization of microtiterplates with integrated optical sensors for oxygen and pH, and their applications to enzyme activity screening, respirometry, and toxicological assays. Sens Actuat B 113:639–648 Walker GM (2000) Yeast physiology and biotechnology. Wiley, West Sussex, pp 144–149 Knorr B, Schlieker H, Hohmann H-P, Weuster-Botz D (2007) Scale-down and parallel operation of the riboflavin production process with Bacillus subtilis. Biochem Eng J 33:263–274 Faust G, Janzen N, Bendig C, Ro¨mer L, Kaufmann K, WeusterBotz D (2014) Feeding strategies enhance high cell density
123
12.
13.
14.
15.
16.
17.
18.
19. 20.
21.
cultivation and protein expression in milliliter-scale bioreactors. Biotechnol J 9:1293–1303 Kusterer A, Krause C, Kaufmann K, Arnold M, Weuster-Botz D (2008) Fully automated single-use stirred-tank bioreactors for parallel microbial cultivations. Bioproc Biosys Eng 31:207–215 Hortsch R, Krispin H, Weuster-Botz D (2011) Process performance of parallel bioreactors for batch cultivation of Streptomyces tendae. Bioproc Biosys Eng 34:297–304 Bendig C, Weuster-Botz D (2013) Reaction engineering analysis of cellulase production with Trichoderma reesei RUT-C30 with intermittent substrate supply. Bioproc Biosys Eng 36:893–900 Riedlberger P, Bru¨ning S, Weuster-Botz D (2013) Characterization of stirrers for screening studies of enzymatic biomass hydrolyses on a milliliter-scale. Bioproc Biosys Eng 36:927–935 Puskeiler R, Kaufmann K, Weuster-Botz D (2005) Development, parallelization and automation of a gas-inducing milliliter-scale bioreactor for high-throughput bioprocess design (HTBD). Biotechnol Bioeng 89:512–523 Weuster-Botz D, Puskeiler R, Kusterer A, Kaufmann K, John G, Arnold M (2005) Methods and milliliter scale devices for highthroughput bioprocess design. Bioproc Biosys Eng 28:109–119 Schmidt M, Weuster-Botz D (2012) Reaction engineering studies of acetone-butanol-ethanol fermentation with Clostridium acetobutylicum. Biotechnol J 7:656–661 Klimant I (2003) Patent US 6,602,716 B1 Monot F, Martin JR, Petitdemange H, Gay R (1982) Acetone and butanol production by Clostridium acetobutylicum in a synthetic medium. Appl Environ Microbiol 44:1318–1324 Korz DJ, Rinas U, Hellmuth K, Sanders EA, Deckwer WD (1995) Simple fed-batch technique for high cell density cultivation of Escherichia coli. J Biotechnol 39:59–65
