An automated workflow for enhancing microbial bioprocess optimization on a novel microbioreactor platform
© Rohe et al.; licensee BioMed Central Ltd. 2012
Received: 10 September 2012
Accepted: 29 October 2012
Published: 31 October 2012
High-throughput methods are widely-used for strain screening effectively resulting in binary information regarding high or low productivity. Nevertheless achieving quantitative and scalable parameters for fast bioprocess development is much more challenging, especially for heterologous protein production. Here, the nature of the foreign protein makes it impossible to predict the, e.g. best expression construct, secretion signal peptide, inductor concentration, induction time, temperature and substrate feed rate in fed-batch operation to name only a few. Therefore, a high number of systematic experiments are necessary to elucidate the best conditions for heterologous expression of each new protein of interest.
To increase the throughput in bioprocess development, we used a microtiter plate based cultivation system (Biolector) which was fully integrated into a liquid-handling platform enclosed in laminar airflow housing. This automated cultivation platform was used for optimization of the secretory production of a cutinase from Fusarium solani pisi with Corynebacterium glutamicum. The online monitoring of biomass, dissolved oxygen and pH in each of the microtiter plate wells enables to trigger sampling or dosing events with the pipetting robot used for a reliable selection of best performing cutinase producers. In addition to this, further automated methods like media optimization and induction profiling were developed and validated. All biological and bioprocess parameters were exclusively optimized at microtiter plate scale and showed perfect scalable results to 1 L and 20 L stirred tank bioreactor scale.
The optimization of heterologous protein expression in microbial systems currently requires extensive testing of biological and bioprocess engineering parameters. This can be efficiently boosted by using a microtiter plate cultivation setup embedded into a liquid-handling system, providing more throughput by parallelization and automation. Due to improved statistics by replicate cultivations, automated downstream analysis, and scalable process information, this setup has superior performance compared to standard microtiter plate cultivation.
KeywordsMicrobioreactor Biolector Bioprocess development Strain screening Heterologous protein expression Automation Liquid-handling
Determination of quantitative process parameters for optimizing production strains and cultivation conditions are key elements during bioprocess development for almost every new microbial product. Especially the optimization of heterologous protein production plays a major role, e.g. in biocatalytic and biopharmaceutical manufacturing processes. The protein expression itself is influenced by many biological and bioprocess parameters, which are normally defined in early development phases and kept constant during the scale-up to production scale. Influencing factors are the choice of the expression host, the promoter as well as its induction, temperature switches, batch or fed-batch mode and different media compositions, to mention only some of them[2, 3]. To optimize a fermentative production process for a new protein, these parameters must be thoroughly chosen, tailored to the target protein and optimized de-novo due to the differences in nature and characteristics of each protein[4–6].
The challenge is that the number of possible parameter combinations exponentially grows with every additional parameter. To give a simple example, 2 parameters with 4 possible values lead to 16 experiments, whereas for 5 parameters in total 1024 experiments are necessary to identify the optimal parameter setting of all combinations. Hence, on the one hand there is a strong demand for significantly increased cultivation throughput in early process development. On the other hand process monitoring and process control should be as high as possible to allow substantial and reliable parameter optimization to ensure scalability of results.
A process optimization in laboratory scale bioreactors providing detailed informative process data is usually accompanied by a low cultivation throughput. In a given timeframe only a few selected process modifications can be tested and the parameter selection process will be often based on empirical knowledge or by trying to adapt a standard process regime for each new heterologous protein. This clearly illustrates that early bioprocess optimization in laboratory scale bioreactors can only provide a limited contribution.
An increase of cultivation throughput by at least two orders of magnitude is achieved quite easily by cultivation in shaking flasks or microtiter plates (MTP)[7, 8]. This is usually accompanied by reduced process information over time and loss of well controlled process conditions present in laboratory scale bioreactors. Several approaches to circumvent those disadvantages were investigated in the last years to hit a middle ground between laboratory scale bioreactors and MTP cultivation[7, 9–12] with so called microbioreactors (MBR). MBR concepts can be differentiated in two approaches: The top-down approach is the miniaturization and parallelization of stirred tank reactors down to 10-100 mL gaining process information easily scaled to laboratory scale bioreactors[13–15]. Beside this, the bottom up approach uses standard formats like MTP or shaking flasks extended with enhanced oxygen transfer, as well as monitoring of main process parameters[17–19]. The miniaturized stirred reactors can be handled automatically e.g. with an integration of the vessels in liquid-handling platforms[15, 20, 21]. With such a setup, repeated sampling during cultivations for offline biomass and product quantification for example in an adjacent spectrophotometer as described by Knorr et al. can be carried out. In contrast to this, a sampling event from MTP based cultivations with volumes below 2 mL is usually a total harvest of the whole reaction volume. In novel MTP based cultivation systems, the online monitoring allows real time quantification of biomass and fluorescence signals, making sampling obsolete for monitoring the expression of fluorescent proteins. Furthermore the achieved online signals can be used to trigger automated pipetting events like addition of initial cell inocula or IPTG solution with automated robotic workstations.
In this contribution, we setup such a smart interactive combination of MTP cultivation system (Biolector) and liquid-handling robotics as a versatile bioprocess optimization platform with high cultivation throughput called Jülich Bioprocess Optimization System (JuBOS). The setup was customized with further hardware components, to enable basic microbial workflows including media preparation, inoculation, sampling and dosing events during cultivation. Further downstream processes like cell separation and photometric protein analysis are possible focusing on the optimization of recombinant protein production.
As a relevant application example, the production of cutinase, a lipase from Fusarium solani pisi was investigated as model enzyme expressed in C. glutamicum, a novel expression host. This Gram-positive, facultative anaerobic bacterium has been successfully developed as production organism for small metabolic intermediates like glutamate and lysine at industrial scale[26, 27]. Furthermore C. glutamicum possesses both the SEC-secretion and the Twin-Arginin-Translocation (TAT) protein secretion pathways, which are known to secrete unfolded and folded proteins, respectively. In addition, no detectable extracellular proteolytic activity is known for this organism, making it suitable for the secretory expression of protease-sensitive proteins. Thus, current research activities deal with the potential of C. glutamicum as producer strain for heterologous proteins[29–32]. The decision about the protein secretion pathway used as well as the secretion efficiency depends strongly on the nature of the signal peptide (SP) fused to the target protein. Unfortunately, the efficiency of the SP is not known a-priori, since it is also determined by the combination of SP and the target protein itself[33, 34]. Computer based tools for prediction of efficient signal peptide sequences provide usually only limited information and were proven to be incorrect. Thus, experimental screening procedures with each protein of interest are essential to rank the SP’s regarding their performance for the secretory production. With every new combination of SP and target protein bioprocess parameters like media components and induction strategy must be optimized de-novo. To meet the requirements of such process optimizations, we describe in this study the validation and application of a MBR system integrated into a robotic platform. In contrast to former publications one challenge is the demand for offline protein quantification of the non-fluorescent cutinase which is met by additional hardware on the automated platform like MTP centrifuge and MTP reader on the JuBOS.
Results and discussion
Sterile environment of the cultivation platform
Biological cross contamination between pipetting events
During sampling procedures and downstream processing operations, the steel needles used for pipetting are in direct contact with cell suspension. To avoid biological cross contaminations, the teflon coated steel needles need to be cleaned and free of CFU for contamination critical steps like automated media preparation or dosing events into the cultivation wells.
For other pipetting procedures which are done in later process phases, like induction with IPTG or sampling events, the strict necessity to avoid any carry-over of cells from one culture to another culture is not prevalently critical. The transfer of at least 10 C. glutamicum cells still present after quick ethanolic flushing will have no effect in the presence of a biomass concentration of at least 10 g.L-1, which corresponds to about 1011 CFU.L-1. This clearly shows the negligible impact of the cross-transferred cells compared to the 1010 times higher cell concentration in the cultivation. Thus, it seems to be a reasonable trade-off between sterility issues and speed to choose the faster variant of cleaning the steel needles if indicated. For automated media preparation this is no choice, since even one single contaminant cell will spoil the experiment.
Automated media preparation with low statistical error
The ability to perform the upstream processing in an automated fashion under sterile conditions allows preparing the cultivation media for the small scale cultivation directly on the robotic deck prior to cultivation. In most applications of parallel small scale cultivations the MTP for cultivation is filled manually instead of doing this automatically[35, 36]. Even for higher throughput this is typically not a limiting step, since it can be achieved quickly using multi-channel pipettes or semi-automated lab dispensers if the same medium is used for all cultivations.
The media preparation becomes a time-critical step, if a diverse set of media compositions has to be prepared, e.g. to investigate the influence of different media components on the production performance of the microbial system. In such cases, the automated media preparation is highly advantageous, since the robotic system 1) shows superior speed in pipetting different media compositions from stock solutions and 2) provides higher reliability of pipetted medium composition accompanied by an electronic documentation of the performed pipetting events.
The automated media preparation with the JuBOS was first validated against the manually prepared medium. Therefore, the minimal medium CG XII for C. glutamicum was prepared by the JANUS workstation from four stock solutions containing single components ((NH4)2SO4, Urea, ZnSO4, CuSO4) in 20 fold concentration and one further stock solution containing all the other components of CG XII medium in twofold concentration. The required volumes of stock solutions were pipetted into the wells of the cultivation plate (i.e. flowerplate) in fourfold replicates and were automatically filled up to the start volume of 950 μL with sterile deionized water. The manually pipetted medium was prepared from the same stock solutions with a final volume of 42 mL and 950 μL was dispensed in four replicates into the Flowerplate. All cultivation wells were inoculated with 50 μL preculture.
To illustrate the benefit of the option for automated media preparation, we applied this method to investigate the effect of two nitrogen sources, namely (NH4)2SO4 and urea on the protein expression of GFP in C. glutamicum. A full factorial design approach was used with variable urea and ammonium sulfate concentration resulting in 48 different media prepared in one MTP by the robotic system. The manual preparation from 4 stock solutions and additional water like described above would take approx. 1.5 hour. This workload is drastically reduced below 10 min by using the automated media preparation proving a fast media preparation method.
The C. glutamicum cells contain plasmid pEKEx2-PhoD-GFP that harbors the gene for a GFP protein that is fused to the signal peptide of the B. subtilis TAT-substrate PhoD. Expression of the PhoD-GFP fusion in C. glutamicum leads to TAT-dependent secretion of the folded GFP protein into the culture supernatant. As comparative signal for protein production, the maximal online fluorescent signal was used and normalized with the maximum biomass signal obtained in each cultivation to achieve a biomass-specific GFP formation (Figure3B). The data indicate that the concentrations of the nitrogen sources in standard CG XII medium (see marker) do not provide optimal conditions for GFP production. The biomass specific fluorescence is 50% higher at reduced ammonium sulfate concentrations (between 0 to 5 g.L-1) and an urea concentration at 10 g.L-1, represented as red area in Figure3B.
With GFP as an example for protein expression in C. glutamicum, the CG XII medium seems to be not ideally suited. In current applications C. glutamicum is mainly used for production of small metabolites from primary metabolism, e.g. lysine, glutamate, valine or succinate and the published cultivation media have been optimized in the past for fast growth and amino acid production[42–44]. But during the last 5 years C. glutamicum gains in importance as a host for secretory protein production. The published cultivation medium used for those protein production processes[30, 32] is also minimal media derived from CG XII and differ mainly in the added trace metal and CaCO3 concentrations. The bulk nitrogen source of both, the CG XII media and the protein secretion media is ammonium sulfate, which shows a negative effect on the fluorescence quantity in this study. Urea appears to be a proper alternative nitrogen source for the production of secretory GFP.
So far, a description of a systematic media optimization for secretory protein production with C. glutamicum has not been investigated. Our results indicate clearly a significant influence of the media composition on the overall production. The optimization of the medium specifically suited for a certain microorganism or target application seems to provide a clear benefit. By the use of the JuBOS such effects can be identified and the automation can be efficiently used for routine media optimization during bioprocess development.
Precise strain characterization by triggered sampling events
The challenge of process related screening experiments in MBR is to increase the experimental throughput while maintaining process relevant conditions and to achieve scalable information about the microbial production characteristics. To keep the throughput in MTP based approaches as high as possible, it is necessary to analyze the characteristic strain parameters in a single cultivation well per clone. A common and easy approach is end-point determination in an “over night cultivation” with a constant cultivation time, synchronous IPTG induction at the beginning and a synchronous harvest of all cultivations wells in the stationary phase[34, 45]. We investigated three cutinase producer strains each with a different signal peptid of SEC-substrates from B. subtilis (Bpr, YmwC, AmyE). The lipolytic activity and its statistical error are analyzed with 9 to 12 fold biological replicates, each with different colonies as starting material for precultures.
Influence of three harvest methods on the lipolytic activity (EA spec ) fused with three different signal peptides (bpr, amyE, ywmC)
A: EAspec [kU.g-1] “Over night cultivation”
B: EAspec [kU.g-1] Triggered harvest
C: EAspec [kU.g-1] Triggered induction + harvest
10.11 ± 2.98
11.84 ± 0.96
13.34 ± 1.25
3.61 ± 1.50
5.31 ± 0.60
4.40 ± 0.54
1.22 ± 0.64
1.62 ± 0.33
1.49 ± 0.24
Mean relative error
Duration of stationary phase
6.0 h ± 1.7 h
0.9 h ± 0.4 h
1.6 h ± 0.3 h
Nevertheless, the advantages of MTP cultivations can be used to compare different strains by synchronizing the biomass growth with enzymatic glucose release media from polysaccharides[48–50]. But up to this time, there is no release medium optimized for C. glutamicum commercially available. A second opportunity is to equalize the growth behavior by an adjustment of the initial cell number based on OD or backscatter values, which can be carried out by automated robotic systems. With this adjustment the authors achieved a synchronous growth pattern during the batch phase. When such a synchronized MTP was sampled, all cultivations were in the same growth phase leading to a fair comparison of different cultivations. This adjustment is reasonable, when the vitality of the starting cell material and the growth rate in the main culture are similar.
In our study, the precultures were inoculated with cell material from single colonies, then cultivated to stationary phase and transferred to the main culture. Thus, both prerequisites for an adjustment of the initial cell material were not given and a strategy to circumvent the additional burden of culture synchronization would be beneficial.
To achieve reliable data from heterologously growing C. glutamicum cultivations, a triggered sampling method was developed on the JuBOS. Depending on the online biomass signal measured by Biolector, each cultivation well was harvested independently 4 h after reaching 4 g.L-1 CDW in the early stationary phase (Figure4, B). Cell suspensions were then stored at 4°C in deep well plates and centrifuged at 4000 g for 10 min. With this strategy, the mean duration of stationary phase was lowered to 0.9 h ± 0.4 h. The mean values of EAspec achieved with biomass triggered sampling events are in the same range of the mean EAspec analyzed with synchronous sampling (Table1). However, the trigger leads to a drastically decreased relative error of 13% compared to 41% which dramatically facilitates the selection of best strains. But this method can be still an unfair comparison, due to the fact, that all cultivations were induced from beginning leading to different induction time for each single cultivation.
To investigate this effect of the induction time, a further opportunity for synchronization of expression pattern was tested. A biomass triggered addition of an inductor substance (e.g. IPTG) leads to a production of recombinant proteins starting always at the same biomass concentration. This method was developed in previous studies and validated with online detectable fluorescent proteins. In our study we combined harvest with this biomass triggered IPTG addition at 4 g.L-1 CDW in a way that IPTG was added at 4 g.L-1 CDW and harvesting was done 4 h later (Figure4C). This was individually done for each cultivation based on the online biomass data fully exploiting the potential of the interactive integration of the cultivation with the automated liquid-handling operation. As it can be seen in Table1 the lipolytic activities of the three constructs show the same ranking like in the other two methods. Interestingly, with this high sophisticated method, a relative error of 13% can be achieved, which is the same error of the method without triggered induction but with induction from beginning. It can be concluded, that synchronization of the growth related harvesting time is more important than synchronization of induction. Reliable data can be achieved by using the triggered sampling strategy with the simple synchronous induction at the beginning of the cultivation. In our case, this method was developed for the screening of signal peptides, which have a significant impact for the development of secretory protein production. Because of the flexible trigger setpoints, this sampling method is also suitable for any other strain characterization or the investigation of different culture conditions.
Profiling of optimal induction conditions
In recombinant protein production processes the induction in later process phases is usually beneficial to avoid inhibiting effects like toxicity or metabolic burden. The main parameters that can be influenced in the cultivation process are the concentration of the inductor and the time point of addition. Simultaneous optimization of both parameters in a microtiter cultivation setup proved to be a fruitful approach during the production of fluorescent proteins. The optimal induction strategy not only depends on the coding plasmid and its promoter, but also on the target protein itself. For this reason, a method for induction profiling, was setup analyzing activity of the cutinase enzyme, which is the heterologous target protein to be secreted in C. glutamicum.
The highest cutinase production of 13.5 ± 1.4 kU.L-1 was observed at low IPTG concentrations (0.2 - 0.3 mmol.L-1) added in the early exponential growth phase. When using the IPTG concentration of 1 mmol.L-1, which is commonly used for lac based promoters, the measured cutinase activity is lowered drastically by nearly 30% from 13.5 kU.L-1 to 9.2 ± 0.9 kU.L-1. Thus, the data show a clear influence of the induction strategy on production behavior of secretory cutinase. Consequentially using this method as an early stage module followed by validation experiments in 1 L bioreactors is strongly recommendable in future bioprocess development for secretory target proteins.
Scalable process data achieved in MTPs
Reliable results for early upstream bioprocess development are normally achieved in stirred tank reactors with well-defined environment. To characterize protein production processes in the MTP cultivation system Biolector, the obtained growth and production behavior has to be validated in comparison to those bioreactors with commonly used scales like 1 L and 20 L. Here, in all scales C. glutamicum producing recombinant cutinase fused to the best performing signal peptide (NprE) of this study was investigated. In 1 mL and 1 L scale two further SPs (YwmC, YpjP) serve as mid and low performance producers to compare the ranking of the SP secretion efficiency. Protein expression was induced at the beginning of the cultivation with 0.5 mM IPTG. Sequential samples were taken every hour, obtaining OD, CDW and lipolytic activity.
In flowerplate cultivations the same preculture, temperature, inoculum size and induction strategy is used. Since no pH regulation can be achieved in MTPs, the medium is buffered with MOPS to an initial pH = 7. The scale-up criterion was a DO > 20% according to Riesenberg et al.. In bioreactors this was achieved by a closed-loop strategy of stirrer and gas flow rate. In flowerplates this criterion was fulfilled at shaking frequency n = 1200 min-1 and filling volume V = 1 mL leading to an OTRMax = 80 mmol.L-1.h-1 (information kindly provided by m2p-labs, Aachen/D).
The progress in MTP based microbioreactor techniques during the last five years has resulted in the easy integration of benchtop instruments with automated liquid-handling stations. The shown validation tests of the Jülich Bioprocess Optimization System (JuBOS) support the usability of such an automated platform for microbial cultivation experiments in high-throughput approach. A complete workflow from sterile media preparation and scalable cultivations to triggered sampling and subsequent at-line analysis of enzyme activity was established for optimization of recombinant cutinase production. Although the cultivation platform was currently used in batch mode only, it is the basis for fed-batch application in the near future employing fed-batch by use of microfluidic technology or slow release systems for substrates.
As a key advantage, the resulting statistical error of productivity data of non fluorescent target protein was drastically decreased from 41% to 13%, due to triggered sampling events to synchronize growth pattern of biological replicates. With these methods it is now possible not only to select production strains in a higher resolution but also to optimize reaction conditions like induction strategies and media components based on reliable results that are directly scalable to stirred tank bioreactors.
The throughput is drastically increased due to 48 parallel cultivation wells of the Biolector. For example, the induction profiling carried out on the JuBOS has a process time of approx. 24 h (6 h preculture, 14 h main culture, 2 h DSP) with a manual workload of approx. 2 h. In comparison to this, 48 induction experiments in a 4 parallel bioreactor system (e.g. DASGIP BioBlock) would have a process time of 312 h (12 runs * [6 h setup time & preculture, 14 h main culture, 6 h DSP & cleaning]). This 13 fold increased throughput might seem an unfair comparison, because usually induction or media optimization experiments are carried out in shaking flasks or standard MTP techniques to increase the throughput. Those approaches however often involve cost intensive iteration steps backwards after failing of scale-up from bench to pilot- or production-scale. With our contribution we can close this gap with a flexible and easy to use automated platform for medium throughput cultivation in the early bioprocess development of fermentative protein production processes.
Strains and media
C. glutamicum ATCC 13032 was used as host strain. Gfp UV gene was ligated to phoD-SP. Cutinase gene from Fusarium solani pisi was ligated to DNA sequences of nprE, ywmC and ypjP SEC- signal peptides. Constructs were cloned into pEKEx2 plasmid using E. coli XL1 blue. Cells were transformed via electroporation, selected on brain heart infusion (BHI) agar plates (37 g.L-1 BHI, 91 g.L-1 sorbitol, 50 mg.L-1 kanamycin) and cultivated in BHI boullion (37 g.L-1 BHI). Cells of exponential growth phase were harvested and incubated in 30% v/v glycerol for 10 min at 20°C and stored in 2 mL aliquots at −80°C. LB agar plates for contamination studies were prepared like described, but without selection marker kanamycin.
Precultures were performed in BHI medium. For all main cultures minimal media CG XII was used with 20 g.L-1 (NH4)2SO4, 5 g.L-1 Urea, 1 g.L-1 KH2PO4, 1 g.L-1 K2HPO4, 13.25 mg.L-1 CaCl2 × 2H2O, 0.25 g.L-1 MgSO4 × 7H2O, 0.2 mg.L-1 biotin, 30 mg.L-1 protocatechuic acid, 10 mg.L-1 FeSO4 × 7H2O, 10 mg.L-1 MnSO4 × H2O, 1 mg.L-1 ZnSO4 × 7H2O, 0.313 mg.L-1 CuSO4 × 5H2O, 0.02 mg.L-1 NiCl2 × 6H2O. The pH was set to 7.0 with 4 M NaOH. In MTP cultivations 42 g.L-1 MOPS was used as buffer reagent. Additionally 25 mg.L-1 kanamycin and up to 0.5 mM IPTG were added. All chemicals were of analytical grade and supplied by Sigma Aldrich.
All MTP cultivations were carried out in 48 well FlowerplatesTM (Art.-No: MTP-48-B, m2p-labs, Aachen/D) incubated in a Biolector device (m2p-labs, Aachen/D). Initial cultivation volume was set to 1 mL per well, including the inoculums of 50 μL. Following cultivation conditions were set in the incubation chamber: shaking diameter 3 mm, frequency 1200 min-1, temperature 30°C and relative humidity 80%. MTPs were sealed with gas permeable membranes (Art.-No: F-GP-10, m2p-labs, Aachen/D) to minimize evaporation. For the given shaking diameter (3 mm) and frequency the manufacturer reported a OTR of 0.08 mol.L-1.h-1 which is very close to values achieved in standard stirred tank bioreactors[55, 56]. Biomass, dissolved oxygen concentration and pH were measured online in each cultivation well with measurement cycle of 10 minutes. Biomass was detected in the Biolector with backscatter measurement at 600 nm and calibrated like described below. Dissolved oxygen concentration as well as pH values were measured by pre- calibrated optodes at the bottom of each well. Precultures for one strain were carried out in 100 mL two baffled shake flasks with 10 mL media. Multi-strain precultures were carried out with 200 μL media in black 96 well microtiter plates (Greiner Bio-One, Frickenhausen/D) were used with 200 μL media, shaken with 900 min-1 and 1.5 mm eccentricity on a Titramax 100 (Biotest, Dreieich/D).
For bioreactor cultivations 2 L DASGIP glass bioreactors (dI = 110 mm, H = 245 mm) equipped with three steel baffles and two six-bladed impellers (d = 45 mm) were used. Dissolved oxygen (DO) was set above 30% with a sequential feedback regulation cascade of stirrer speed (800–1500 rpm) and gas flow (60 L.h-1 and 120 L.h-1). Offgas content of CO2 and O2 was measured with gas analysis (DASGIP, Juelich/D). The pH was set to 7 and controlled with 4 M HCl and 4 M NaOH. Starting volume for all processes was 1 L. Addition of anti foam was controlled by a foam probe. Every hour 5 mL was drawn. Calibration of the peristaltic pumps and recording and controlling of process parameters was done with the process control software DASGIP Manager (DASGIP, Juelich/D).
For scale- up experimentation a 30 L Chemap bioreactor (dI = 260 mm, H = 560 mm) equipped with two six bladed impellers (d = 90 mm) was used. A process control system (iCCC 2000, PCS Systemtechnik GmbH, Munich/ D) enables pH control and DO control with a cascade of stirrer speed (400 min-1 to 950 min-1) , gas flow rate between 850 L.h- 1 to 1800 L.h- 1 and up to 0.8 bar overload pressure. Inoculation was carried out in all scales with 5% v/v vital preculture. 20 mL samples were drawn every hour to quantify biomass, substrate and product concentration.
The detection of the lipolytic activity of cutinase was carried out with the p-nitrophenyl-palmitate (pNPP) Assay according to[34, 57]. The final concentration of the substrate was 0.8 mM. Optical density (λ = 450 nm) was detected in a Genios Platereader (TECAN, Crailsheim/D) every 25 second for 40 minutes with 3 seconds of shaking before every measurement. Lipolytic activity [kU.L-1 was determined from values of OD450min-1. Samples were diluted with deionized water till values of OD.min-1 > 0.1. Samples with water instead of enzyme solution were used as blanks. The enzymatic activity was calculated using a molar extinction coefficient of 15 cm2.μmol-1 according to.
Biomass quantification and backscatter calibration
For the determination of cell dry weight (CDW) 1 mL cell suspension was centrifuged at 13000 g for 10 min in predryed and weighed 1.5 mL eppendorf cups. Supernatant was discarded and after pellet resuspended in 1 mL 0.9% w/v NaCl solution and centrifuged again at 13000 g for 10 minutes. After drying at 80°C for 18 h the mass difference was used to calculate CDW. For measurement of OD600, samples were diluted with 0.9% NaCl below OD600 = 0.8 measured in a Pharmspec UV 1700 photometer (Shimadzu, Duisburg/D).
Backscatter measurement was calibrated in Flowerplates sealed with a gas-permeable membrane (m2p-labs, Aachen/D) at V = 1000 μL, gain = 20, shaking speed 1200 rpm. The backscatter calibration analysis was done with a dilution series of C. glutamicum biomass suspension in a range of 0 to 20 g.L-1 CDW leading to the linear regression equation of CDW = 0.48*BS g.L-1.au-1 - 0.78 g.L-1 with R2 > 0.98. The calibration was confirmed with samples of 1 L batch bioreactors, whose backscatter and CDW were measured like described above.
Robotic platform for microbioreactor cultivation
Core instrument of the platform is the liquid-handling station JANUS Integrator with a 400 mm overhanging extension (PerkinElmer, Waltham MA/USA) in two arm configuration consisting of a robotic pipetting arm (VarispanTM) equipped with 8 steel needles for liquid transfer up to 1.2 mL and a gripper arm for MTP transport on the platform. Liquid-handling procedures are conducted by pumping deionized water as a system liquid through the tube system attached to the steel needles. The system liquid can be switched by an automated 3 way valve (Bürkert, Ingelfingen/D) to 70% v/v ethanol for cleaning procedures to avoid biological cross contamination. The MTP based cultivation system Biolector (m2p-labs, Aachen/D) is installed on the robotic deck in a way that the MTP plate is fully accessible, allowing automated transfer of liquid, i.e. dosage or sampling, as well as replacement of the complete MTP plate. The whole setup is installed in a laminar flow hood (Cleanair, Woerden/NL) with a high-efficiency-particulate-air- filter (HEPA-filter) generating sterile environment on the robotic deck. Program automation protocols and methods were written in the command software WinPrep (Version 4.6, PerkinElmer) to manage the required actions of the liquid-handling robotic system. The software Robolector agent V.1.2 (m2p-labs, Aachen/D) organizes the communication between WinPrep software and the Biolector. Cultivation data measured online by the Biolector (biomass, pH, DO) can be used as trigger signals to automatically initiate pipetting events like sampling and dosing for each single cultivation well independently.
The sampled cell suspension is stored in a deep well plate (DWP) placed on a cooling rack (MeCour, Groveland MA/USA) connected with a Cryostat Unichiller (Huber, Offenburg/D) in a temperature range beween - 10°C and + 20°C on the robotic deck. A MTP centrifuge (IXION, Sias/CH) is also fully accessible for the robotic system. Supernatant was separated in DWPs via centrifugation at 4000 g for 15 minutes.
Sterility test of steel needles
1000 μL of exponential growing cell suspension (OD600 = 5) are aspirated and released again. Then different treatments like washing with water, EtOH and incubating with EtOH were investigated along with positive control without any washing. Afterwards 50 μL of the upcoming system liquid was pipetted into 1000 μL VE- Water and CFU was determined on LB agar plates.
Cleaning in place
Cell dry weight
Colony forming units
Green fluorescent protein
Sterilizing in place
The financial support from Bio.NRW, a funding initiative of the local government of North Rhine-Westphalia is gratefully acknowledged (Ziel2, grant No W0805wb003d).
- Doig SB F, Lye GL: High throughput screening and process optimisation. Basic Biotechnology. Edited by: Ratledge CK B. 2006, 3rdGoogle Scholar
- Baneyx F: Recombinant protein expression in Escherichia coli. Curr Opin Biotech. 1999, 10: 411-421. 10.1016/S0958-1669(99)00003-8.View ArticleGoogle Scholar
- Weickert MJ, Doherty DH, Best EA, Olins PO: Optimization of heterologous protein production in Escherichia coli. Curr Opin Biotech. 1996, 7: 494-499. 10.1016/S0958-1669(96)80051-6.View ArticleGoogle Scholar
- Hust M, Steinwand M, Al-Halabi L, Helmsing S, Schirrmann T, Dubel S: Improved microtitre plate production of single chain Fv fragments in Escherichia coli. New Biotechnol. 2009, 25: 424-428. 10.1016/j.nbt.2009.03.004.View ArticleGoogle Scholar
- Donovan RS, Robinson CW, Glick BR: Review: optimizing inducer and culture conditions for expression of foreign proteins under the control of the lac promoter. J Ind Microbiol. 1996, 16: 145-154. 10.1007/BF01569997.View ArticleGoogle Scholar
- Makrides SC: Strategies for achieving high-level expression of genes in Escherichia coli. Microbiol Rev. 1996, 60: 512-538.Google Scholar
- Duetz WA: Microtiter plates as mini-bioreactors: miniaturization of fermentation methods. Trends Microbiol. 2007, 15: 469-475. 10.1016/j.tim.2007.09.004.View ArticleGoogle Scholar
- Schapper D, Alam MN, Szita N, Eliasson Lantz A, Gernaey KV: Application of microbioreactors in fermentation process development: a review. Anal Bioanal Chem. 2009, 395: 679-695. 10.1007/s00216-009-2955-x.View ArticleGoogle Scholar
- Hermann R, Lehmann M, Buchs J: Characterization of gas–liquid mass transfer phenomena in microtiter plates. Biotechnol Bioeng. 2003, 81: 178-186. 10.1002/bit.10456.View ArticleGoogle Scholar
- Betts JI, Baganz F: Miniature bioreactors: current practices and future opportunities. Microb Cell Fact. 2006, 5: 21-10.1186/1475-2859-5-21.View ArticleGoogle Scholar
- Bareither R, Pollard D: A Review of Advanced Small-Scale Parallel Bioreactor Technology for Accelerated Process Development: Current State and Future Need. Biotechnol Progr. 2011, 27: 2-14. 10.1002/btpr.522.View ArticleGoogle Scholar
- Jourdier E, Poughon L, Larroche C, Monot F, Chaabane FB: A new stoichiometric miniaturization strategy for screening of industrial microbial strains: application to cellulase hyper-producing Trichoderma reesei strains. Microb Cell Fact. 2012, 11: 70-10.1186/1475-2859-11-70.View ArticleGoogle Scholar
- Ge X, Hanson M, Shen H, Kostov Y, Brorson KA, Frey DD, Moreira AR, Rao G: Validation of an optical sensor-based high-throughput bioreactor system for mammalian cell culture. J Biotechnol. 2006, 122: 293-306. 10.1016/j.jbiotec.2005.12.009.View ArticleGoogle Scholar
- Puskeiler R, Kaufmann K, Weuster-Botz D: Development, parallelization, and automation of a gas-inducing milliliter-scale bioreactor for high-throughput bioprocess design (HTBD). Biotechnol Bioeng. 2005, 89: 512-523. 10.1002/bit.20352.View ArticleGoogle Scholar
- Ahn WS, Ahn JY, Jung CH, Hwang KY, Kim EE, Kim J, Im H, Kim JO, Yu MH, Lee C: Optimization of expression conditions for soluble protein by using a robotic system of multi-culture vessels. J Microbiol Biotechnol. 2007, 17: 1868-1874.Google Scholar
- Funke M, Diederichs S, Kensy F, Muller C, Buchs J: The baffled microtiter plate: increased oxygen transfer and improved online monitoring in small scale fermentations. Biotechnol Bioeng. 2009, 103: 1118-1128. 10.1002/bit.22341.View ArticleGoogle Scholar
- Anderlei T, Zang W, Buchs J: Online respiration activity measurement (OTR, CTR, RQ) in shake flasks. Biochem Eng J. 2004, 17: 187-194. 10.1016/S1369-703X(03)00181-5.View ArticleGoogle Scholar
- Wittmann C, Kim HM, John G, Heinzle E: Characterization and application of an optical sensor for quantification of dissolved O2 in shake-flasks. Biotechnol Lett. 2003, 25: 377-380. 10.1023/A:1022402212537.View ArticleGoogle Scholar
- Kensy F, Zimmermann HF, Knabben I, Anderlei T, Trauthwein H, Dingerdissen U, Buchs J: Oxygen transfer phenomena in 48-well microtiter plates: Determination by optical monitoring of sulfite oxidation and verification by real-time measurement during microbial growth. Biotechnol Bioeng. 2005, 89: 698-708. 10.1002/bit.20373.View ArticleGoogle Scholar
- Puskeiler R, Kusterer A, John GT, Weuster-Botz D: Miniature bioreactors for automated high-throughput bioprocess design (HTBD): reproducibility of parallel fed-batch cultivations with Escherichia coli. Biotechnol Appl Biochem. 2005, 42: 227-235. 10.1042/BA20040197.View ArticleGoogle Scholar
- Huber R, Ritter D, Hering T, Hillmer AK, Kensy F, Muller C, Wang L, Buchs J: Robo-Lector - a novel platform for automated high-throughput cultivations in microtiter plates with high information content. Microb Cell Fact. 2009, 8: 42-10.1186/1475-2859-8-42.View ArticleGoogle Scholar
- Knorr B, Schlieker H, Hohmann H-P, Weuster-Botz D: Scale-down and parallel operation of the riboflavin production process with Bacillus subtilis. Biochem Eng J. 2007, 33: 263-274. 10.1016/j.bej.2006.10.023.View ArticleGoogle Scholar
- Zimmermann HF, Degussa JR: A Fully Automated Robotic System for High Throughput Fermentation. J Assoc Lab Automation. 2006, 11: 134-137. 10.1016/j.jala.2006.03.003.View ArticleGoogle Scholar
- Samorski M, Muller-Newen G, Buchs J: Quasi-continuous combined scattered light and fluorescence measurements: a novel measurement technique for shaken microtiter plates. Biotechnol Bioeng. 2005, 92: 61-68. 10.1002/bit.20573.View ArticleGoogle Scholar
- Kensy F, Zang E, Faulhammer C, Tan RK, Buchs J: Validation of a high-throughput fermentation system based on online monitoring of biomass and fluorescence in continuously shaken microtiter plates. Microb Cell Fact. 2009, 8: 31-10.1186/1475-2859-8-31.View ArticleGoogle Scholar
- Park JH, Lee SY: Metabolic pathways and fermentative production of L-aspartate family amino acids. Biotechnol J. 2010, 5: 560-577. 10.1002/biot.201000032.View ArticleGoogle Scholar
- Eggeling L, Bott M: Handbook of Corynebacterium glutamicum. 2005, Boca Raton, FL: CRC PressView ArticleGoogle Scholar
- Billman-Jacobe H, Wang L, Kortt A, Stewart D, Radford A: Expression and secretion of heterologous proteases by Corynebacterium glutamicum. Appl Environ Microb. 1995, 61: 1610-1613.Google Scholar
- Meissner D, Vollstedt A, van Dijl J, Freudl R: Comparative analysis of twin-arginine (Tat)-dependent protein secretion of a heterologous model protein (GFP) in three different Gram-positive bacteria. Appl Microbiol Biot. 2007, 76: 633-642. 10.1007/s00253-007-0934-8.View ArticleGoogle Scholar
- Teramoto H, Watanabe K, Suzuki N, Inui M, Yukawa H: High yield secretion of heterologous proteins in Corynebacterium glutamicum using its own Tat-type signal sequence. Appl Microbiol Biot. 2011, 91: 677-687. 10.1007/s00253-011-3281-8.View ArticleGoogle Scholar
- Kikuchi Y, Itaya H, Date M, Matsui K, Wu L-F: Production of Chryseobacterium proteolyticum protein-glutaminase using the twin-arginine translocation pathway in Corynebacterium glutamicum. Appl Microbiol Biot. 2008, 78: 67-74. 10.1007/s00253-007-1283-3.View ArticleGoogle Scholar
- Date M, Itaya H, Matsui H, Kikuchi Y: Secretion of human epidermal growth factor by Corynebacterium glutamicum. Lett Appl Microbiol. 2006, 42: 66-70. 10.1111/j.1472-765X.2005.01802.x.View ArticleGoogle Scholar
- Degering C, Eggert T, Puls M, Bongaerts J, Evers S, Maurer KH, Jaeger KE: Optimization of Protease Secretion in Bacillus subtilis and Bacillus licheniformis by Screening of Homologous and Heterologous Signal Peptides. Appl Environ Microb. 2010, 76: 6370-6376. 10.1128/AEM.01146-10.View ArticleGoogle Scholar
- Brockmeier U, Caspers M, Freudl R, Jockwer A, Noll T, Eggert T: Systematic screening of all signal peptides from Bacillus subtilis: a powerful strategy in optimizing heterologous protein secretion in Gram-positive bacteria. J Mol Biol. 2006, 362: 393-402. 10.1016/j.jmb.2006.07.034.View ArticleGoogle Scholar
- David F, Westphal R, Bunk B, Jahn D, Franco-Lara E: Optimization of antibody fragment production in Bacillus megaterium: The role of metal ions on protein secretion. J Biotechnol. 2010, 150: 115-124.View ArticleGoogle Scholar
- Kottmeier K, Muller C, Huber R, Buchs J: Increased product formation induced by a directed secondary substrate limitation in a batch Hansenula polymorpha culture. Appl Microbiol Biot. 2010, 86: 93-101. 10.1007/s00253-009-2285-0.View ArticleGoogle Scholar
- Nealon AJ, Willson KE, Pickering SC, Clayton TM, O'Kennedy RD, Titchener-Hooker NJ, Lye GJ: Use of operating windows in the assessment of integrated robotic systems for the measurement of bioprocess kinetics. Biotechnol Progr. 2005, 21: 283-291. 10.1021/bp049868+.View ArticleGoogle Scholar
- van Ooyen J, Noack S, Bott M, Reth A, Eggeling L: Improved L-lysine production with Corynebacterium glutamicum and systemic insight into citrate synthase flux and activity. Biotechnol Bioeng. 2012, 109: 2070-2081. 10.1002/bit.24486.View ArticleGoogle Scholar
- Gourdon P, Lindley ND: Metabolic analysis of glutamate production by Corynebacterium glutamicum. Metab Eng. 1999, 1: 224-231. 10.1006/mben.1999.0122.View ArticleGoogle Scholar
- Blombach B, Schreiner M, Bartek T, Oldiges M, Eikmanns B: Corynebacterium glutamicum tailored for high-yield L-valine production. Appl Microbiol Biot. 2008, 79: 471-479. 10.1007/s00253-008-1444-z.View ArticleGoogle Scholar
- Litsanov B, Kabus A, Brocker M, Bott M: Efficient aerobic succinate production from glucose in minimal medium with Corynebacterium glutamicum. Microb Biotechnol. 2012, 5: 116-128. 10.1111/j.1751-7915.2011.00310.x.View ArticleGoogle Scholar
- Weuster-Botz D, Kelle R, Frantzen M, Wandrey C: Substrate Controlled Fed-Batch Production of l-Lysine with Corynebacterium glutamicum. Biotechnol Progr. 1997, 13: 387-393. 10.1021/bp970034j.View ArticleGoogle Scholar
- Keilhauer C, Eggeling L, Sahm H: Isoleucine synthesis in Corynebacterium glutamicum: molecular analysis of the ilvB-ilvN-ilvC operon. J Bacteriol. 1993, 175: 5595-5603.Google Scholar
- von der Osten CH, Gioannetti C, Sinskey AJ: Design of a defined medium for growth of Corynebacterium glutamicum in which citrate facilitates iron uptake. Biotechnol Lett. 1989, 11: 11-16. 10.1007/BF01026778.View ArticleGoogle Scholar
- Hallborn J, Carlsson R: Automated screening procedure for high-throughput generation of antibody fragments. Biotechniques. 2002, 33: 30-37.Google Scholar
- Kondragunta B, Drew JL, Brorson KA, Moreira AR, Rao G: Advances in Clone Selection Using High-Throughput Bioreactors. Biotechnol Progr. 2010, 26: 1095-1103.Google Scholar
- Wenk P, Hemmerich J, Müller C, Kensy F: High-Throughput Bioprocess Development in Shaken Microbioreactors. Chem-Ing-Tech. 2012, 84: 704-714. 10.1002/cite.201100206.View ArticleGoogle Scholar
- Huber R, Scheidle M, Dittrich B, Klee D, Buchs J: Equalizing growth in high-throughput small scale cultivations via precultures operated in fed-batch mode. Biotechnol Bioeng. 2009, 103: 1095-1102. 10.1002/bit.22349.View ArticleGoogle Scholar
- Panula-Perala J, Siurkus J, Vasala A, Wilmanowski R, Casteleijn MG, Neubauer P: Enzyme controlled glucose auto-delivery for high cell density cultivations in microplates and shake flasks. Microb Cell Fact. 2008, 7: 31-10.1186/1475-2859-7-31.View ArticleGoogle Scholar
- Siurkus J, Panula-Perala J, Horn U, Kraft M, Rimseliene R, Neubauer P: Novel approach of high cell density recombinant bioprocess development: Optimisation and scale-up from microlitre to pilot scales while maintaining the fed-batch cultivation mode of E. coli cultures. Microb Cell Fact. 2010, 9: 35-10.1186/1475-2859-9-35.View ArticleGoogle Scholar
- Huber R, Palmen TG, Ryk N, Hillmer AK, Luft K, Kensy F, Buchs J: Replication methods and tools in high-throughput cultivation processes - recognizing potential variations of growth and product formation by on-line monitoring. BMC Biotechnol. 2010, 10: 22-10.1186/1472-6750-10-22.View ArticleGoogle Scholar
- Riesenberg D, Menzel K, Schulz V, Schumann K, Veith G, Zuber G, Knorre WA: High cell density fermentation of recombinant Escherichia coli expressing human interferon alpha 1. Appl Microbiol Biot. 1990, 34: 77-82.View ArticleGoogle Scholar
- Funke M, Buchenauer A, Schnakenberg U, Mokwa W, Diederichs S, Mertens A, Muller C, Kensy F, Buchs J: Microfluidic BioLector-Microfluidic Bioprocess Control in Microtiter Plates. Biotechnol Bioeng. 2010, 107: 497-505. 10.1002/bit.22825.View ArticleGoogle Scholar
- Eikmanns BJ, Kleinertz E, Liebl W, Sahm H: A family of Corynebacterium glutamicum/Escherichia coli shuttle vectors for cloning, controlled gene expression, and promoter probing. Gene. 1991, 102: 93-98. 10.1016/0378-1119(91)90545-M.View ArticleGoogle Scholar
- Yawalkar AA, Heesink ABM, Versteeg GF, Pangarkar VG: Gas–liquid Mass Transfer Coefficient in Stirred Tank Reactors. Can J Chem Eng. 2008, 80: 840-848. 10.1002/cjce.5450800507.View ArticleGoogle Scholar
- Van‘t Riet K: Mass transfer in fermentation. Trends Biotechnol. 1983, 1: 113-119. 10.1016/0167-7799(83)90034-3.View ArticleGoogle Scholar
- Winkler UK, Stuckmann M: Glycogen, hyaluronate, and some other polysaccharides greatly enhance the formation of exolipase by Serratia marcescens. J Bacteriol. 1979, 138: 663-670.Google Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.