# Analysis of the influence of flame sterilization included in sampling operations on shake-flask cultures of microorganisms

## Abstract

Shake-flask cultures of microorganisms involve flame sterilization during sampling, which produces combustion gas with high CO2 concentrations. The gaseous destination has not been deeply analyzed. Our aim was to investigate the effect of flame sterilization on the headspace of the flask and on the shake-flask culture. In this study, the headspace CO2 concentration was found to increase during flame sterilization ~0.5–2.0% over 5–20 s empirically using the Circulation Direct Monitoring and Sampling System. This CO2 accumulation was confirmed theoretically using Computational Fluid Dynamics; it was 9% topically. To evaluate the influence of CO2 accumulation without interference from other sampling factors, the flask gas phase formed by flame sterilization was reproduced by aseptically supplying 99.8% CO2 into the headspace, without sampling. We developed a unit that can be sampled in situ without interruption of shaking, movement to a clean bench, opening of the culture-plug, and flame sterilization. We observed that the growth behaviour of Escherichia coli, Pelomonas saccharophila, Acetobacter pasteurianus, and Saccharomyces cerevisiae was different depending on the CO2 aeration conditions. These results are expected to contribute to improving microbial cell culture systems.

## Introduction

Rotary shake-flask culturing was first used for the cultivation of Aspergillus flavus in 19331, and is now widely used for the cultivation of various microorganisms. Shake-flask culturing is different from surface culturing where the cells are directly exposed to high oxygen concentrations. In shake-flask cultures, the cells are exposed to low oxygen concentrations since the microorganisms are suspended in the culture broth. Therefore, a lot of work has been done to improve oxygen supply to shake-flask cultures. For example, a shouldered, flat-bottom Kolben (termed a Sakaguchi flask) for reciprocating, shaking cultures was developed for submerged cultures of Aspergillus2. In recent years, new shaking methods using resonant acoustic mixing with ~6 times higher kLa have been reported3.

Presently, monitoring devices for shake-flask cultures are not yet common. Flask shaking is often interrupted in order to investigate changes in various culture parameters with time and the sampling procedure has no established restrictions, except for the prevention of contamination. This process is dependent on the skills of the experimenter. A temporary decrease in oxygen transfer rate has been reported when the flask is taken from the shaker to a laminar flow bench to perform sampling11,14. Moreover, it has been reported that, during shake-flask culturing of soil samples, different structures of microbial communities are formed depending on sampling and intermittent opening of the culture-plug for as little as 30 s15. Some cultivation experts may also be aware of the effects of subtle manipulations based on empirical knowledge. However, except for the above, there is currently no report on the influence of the sampling operation on shake-flask cultures. Thus, there is a paucity of information on the influence of conventional sampling procedures on culture parameters. We, therefore, investigated other possible effects of the sampling operation (other than the interruption of shaking and opening of the culture-plug) on microorganism growth.

We have reported previously that the changes in CO2 concentration in the flask gas phase during shaking culture affect the structure of the cultured microbial community15. Among the various operations in standard sampling, it is presumed that those causing changes to the gas environment of the flask (e.g., flame sterilization or opening the culture-plug) have significant effects on the microorganisms in the culture broth. However, it is difficult to evaluate the direct influence of flame sterilization and culture-plug opening on the culture, since these procedures are complicated manual operations that are accompanied by interruption of shaking. Reportedly, with computational fluid dynamics (CFD) analysis, there is a high concentration of CO2 around the Bunsen burner flame due to the combustion of methane16. Based on this information, we hypothesized that the high concentration of CO2 generated at the flask mouth during flame sterilization would move into the flask headspace. Therefore, we designed experimental systems to test this hypothesis. The present study combined CFD analysis of the flame sterilization operation, CDMSS5, and a system that can aerate the flask gas phase (termed Automatic Aeration Flask System (AAFS))15. The aim was to analyze the effect of flame sterilization on the headspace of the flask and on the shake-flask culture.

## Results

### CO2 behavior in flame sterilization

A CFD analysis was also performed in order to understand, in more detail, the effect of the flame sterilization procedure. Figure 4 shows the distributions of (a) CO2 concentration and (b) temperature on a vertical cross section through the center of the flask from 1–20 s. The CO2 concentration inside the flask increased with time. The temperature exceeded 500 °C in the upper portions of the flask, but remained ambient in other portions. Near the inner wall of the flask, there were regions where the CO2 concentration was high and the temperature was low. This suggests that water vapor in the burned gas condensed. It should be noted that condensed water, which includes CO2 and nitrogen oxides and is acidic (pH ~ 3), is produced if the flask’s surface becomes foggy. Supplementary Figs. S2 and S3 online show the time-dependent changes in CO2 concentration and temperature along the center axis of the flask. As mentioned above, CO2 inflows as time passes, reaching a maximum concentration of approximately 9%, which is lower than the concentration when methane undergoes stoichiometric combustion. In contrast, the temperature remains nearly constant after t = 2 s, which indicates that the culture fluids would not be heated by this process.

Regarding the flow of burned gas inside the flask, Fig. 5 shows the streamlines from the burner inlet at t = 1 and 20 s. The midline cross section is shown for clarity and the line colors denote the magnitude of velocity. Incidentally, when the flask is placed in the flame at t = 0 s, it becomes a barrier to the flame’s flow. Clockwise eddies are generated downstream. The shape of the eddies essentially does not change between t = 1 and 20 s, but the velocity does increase. Because the burned gas possesses increased temperature and buoyancy, it inflows along the upper part of the flask. Then, the gas is cooled, and CO2 and H2O flow inside the flask by diffusion and advection.

### Effect of intermittent CO2 changes on microorganisms during shake-flask culture

Considering the conventional sampling process of the shake-flask culture (Fig. 1), when investigating the influence of CO2 accumulation in the gas phase of the flask due to flame sterilization, the interruption of shaking and opening the culture-plug will also affect the culture. Thus, separating the effects of flame sterilization from other effects is not possible using the conventional method. Therefore, the AAFS15, which can introduce fresh air into the flask aseptically without interrupting the culturing, was modified as shown in Supplementary Fig. S4 online to allow intermittent aeration with 99.8% (v/v) CO2. At a CO2 aeration rate of 30 mL/min, the increase in CO2 concentration was the same as that obtained during flame sterilization at an inclination angle of 25 ° (see Supplementary Fig. S5 online). The maximum CO2 concentration increased in proportion to aeration time. Using a 500 mL Erlenmeyer flask aerated with 99.8% (v/v) CO2 at 30 mL/min, as the aeration time increased by 1 s, the CO2 concentration increased by ~0.1% (v/v) (see Supplementary Fig. S5 online).

To investigate the effect of unintentional CO2 accumulation in shake-flask cultures of Escherichia coli and Pelomonas saccharophila, each time sampling was performed without interruption of shaking and then 99.8% (v/v) CO2 was aseptically aerated into the gas phase of the flask for 90 and 30 s at 30 mL/min using AAFS. This sampling and intermittent aeration occurred after 2, 4, 6, 8, 10, 12, 15, 25, and 35 h of culturing. Compared with that in the corresponding control cultures (without aeration), the pH was different in the E. coli cultures, while the U.O.D.600 was different in the P. saccharophila cultures (Fig. 6a,b). Intermittent aeration of E. coli cultures with CO2 by AAFS had no effect on the U.O.D.600 time course (Fig. 6a). However, the pH of the E. coli culture decreased similarly under all the conditions until 10 h. The pH of the control culture increased sharply between 10 and 15 h, but for conditions a and b (90 and 30 s aeration, respectively), the pH increase occurred between 15 and 25 h. The pH values between 25 and 35 h were almost the same under all conditions (Fig. 6a).

Intermittent aeration of P. saccharophila cultures with CO2 by AAFS had no effect on the pH time course. Under all conditions, the pH decreased gradually to ~4.5 (Fig. 6b). The U.O.D.600 of P. saccharophila cultured under a and b conditions (90 and 30 s aeration, respectively) gradually increased up to 6 h, similar to that of the control (Fig. 6b). Condition b produced a U.O.D.600 value that was higher than that of the control between 8 and 15 h (considered the stationary phase), which subsequently converged with the control value (Fig. 6b, square). In the case of condition a, the growth phase was prolonged by 2 h compared to that of the control; the U.O.D.600 rose at ~10 h, and then a slight decrease was observed (Fig. 6b, triangle). However, the U.O.D.600 remained higher than in the other conditions (control and a). The U.O.D.600 values at 10 h were approximately 2.38 (control), 3.22 (condition b), and 4.00 (condition a).Thus, the U.O.D.600 obtained with condition a was 1.68 times higher than that of the control. In the dead phase (35 h), the U.O.D.600 values were 2.10 (control), 2.22 (condition b), and 3.06 (condition a), and the value with condition a was 1.46 times higher than that of the control. The same culture experiments were performed using A. pasteurianus and S. cerevisiae as aerobic microorganisms (Fig. 6c,d). As compared with the control without CO2 aeration by AAFS, there was almost no change in pH change during the time course of both A. pasteurianus and S. cerevisiae (Fig. 6c,d). In the case of U.O.D.600, A. pasteurianus marked differences, but S. cerevisiae showed little change. The U.O.D.600 of A. pasteurianus cultured under a and b conditions (90 and 30 s aeration, respectively) gradually increased up to 15 h, similar to that of the control. Thereafter, the values of U.O.D.600 under these conditions (a and b) rapidly increased until 35 h and reached the stationary phase. The control continued to increase slowly until 55 h. The U.O.D.600 values at 35 h were approximately 1.59 (control), 2.06 (condition a), and 2.71 (condition b) (Fig. 6c). Thus, the U.O.D.600 obtained with condition b was 1.7 times higher than that of the control.

## Discussion

As the flame sterilization time shortened, the CO2 concentration in the headspace of the flask lowered, and the influence on the culture lessened. In order to perform flame sterilization with minimal impact on the culture, it may be preferred to heat the neck located below the mouth of the flask instead of opening the mouth for a very short time. However, the Bunsen burner flame near the mouth of the flask creates a temporary negative pressure, and combustion gas containing high CO2 concentrations near the burner flows into the flask headspace. This phenomenon cannot be avoided during the flame sterilization operation. Flame sterilization using a Bunsen burner performs two roles: 1, sterilizing microorganisms attached to the flask mouth and culture-plug surface due to high temperatures; 2, blowing away attached microorganisms and dust due to the high airflow. This flame sterilization operation is a universal method used during microbial cultivation, not only for flask cultures but also for test tube slant cultures. It is also performed at the time of subculturing and inoculation. Currently, there is no method that accomplishes both points 1 and 2 above without the accumulation of CO2. High-speed hot air treatment can be considered as an alternative to flame sterilization. However, hot air is more difficult to see than a flame.

It has been reported that standard sampling of shake-flask cultures involving the cessation of shaking leads to sharp decreases in oxygen transfer rate during sampling11,14. Opening the culture-plug also leads to unplanned ventilation of the gas phase in the flask15. The present study has shown that the flask gas phase is altered by the flame sterilization operation. In order to perform shake-flask culturing with good reproducibility, it is necessary to prepare the flasks for sampling5 and standardize the conventional sampling operations. Furthermore, it is necessary to construct a system that permits sampling without interrupting the shaking. Sampling techniques for built-in bioreactors, which cannot be transferred to clean benches, have been established, and by applying the same techniques to shake flasks, new sampling methods are expected to be developed in the future.

Balancing pH and CO2 in culture experiments is very critical. When dissolved in culture medium, CO2 exists as three forms: CO32−, HCO3, and CO2. In shaking cultures of E. coli, the pH decreased during the first 10 h, but then increased; thus, the pH fluctuated between 5.8 and 8.7. In the pH range of 5.8 to 6.8, CO32− is hardly present and carbon dioxide exists mainly as HCO3 and CO217. In particular, when the pH is higher than ~6.3, HCO3 becomes the main component and tends to buffer the pH. Accordingly, when CO2 was intermittently aerated by AAFS, the pH was maintained at ~6 within the first 10 h and was comparable to the value obtained in the control cultures (no aeration).The growth of E. coli was minimally affected because the pH was within its growth range18. It was suggested that the same applies to S. cerevisiae whose growth was not changed by intermittent CO2 aeration. Regarding the growth inhibition of S. cerevisiae, it was reported that the CO2 concentration did not change up to 50% in the culture using the reactor19. In the shake-flask culture of P. saccharophila, the pH gradually decreased from 6.7 at the start of culturing to 4.6 at the end. Even in this pH range, the carbon dioxide existed mainly as HCO3 and CO2 because CO32− is mainly present under strongly alkaline conditions. This culture was more acidic than the E. coli culture, and especially below pH 6.3, CO2 is the main component. Certainly, the pH fluctuation was smaller than that of the control, although it was only slight (around pH 6) during the first 6–10 h of culture. After 12 h of culture, buffering by CO2 aeration was not observed, but growth under intermittent CO2 aeration was improved compared with that of the controls. This was probably because the dissolved CO2 concentration was maintained at high levels, and not because of pH buffering capacity. It seemed that the growth of A. pasteurianus increased by intermittent CO2 aeration like the growth of P. saccharophila. It should also be noted that when the surface of the flask mouth becomes cloudy during flame sterilization, acidic (pH ~ 3) condensed water is generated, containing CO2 and nitrogen oxides.

## Methods

### Analysis of manual flame sterilization in sampling process using CDMSS

Manual flame sterilization was performed using a Bunsen burner, 500 mL Erlenmeyer flask with small pipes on the sides, and CDMSS. During the flame sterilization process, the gas in the headspace of the flask was monitored using CDMSS by circulation at 50 mL/min, and the maximum CO2 concentration was recorded. The manual flame sterilization was performed for 5, 10, and 20 s; each time point was repeated four or more times, and the average value was calculated. All experiments were performed at room temperature (approximately 25 °C).

### Analysis of flame sterilization operation with fixed operating angle and time using CDMSS

The system shown in Supplementary Fig. S1 online was constructed, and the effect of flame sterilization on the CO2 concentration in the headspace of the flask was analyzed. Although the flask was brought near the Bunsen burner in the manual experiments, in this system, the position and angle of the flask were fixed when the Bunsen burner was brought close to the flask. After the operation time, the slide plate was moved to separate the Bunsen burner from the flask while monitoring the headspace of the flask using CDMSS (see Supplementary Fig. S1 online, Figs. 16), and the maximum CO2 concentration was recorded. The operation time and inclination angle were set with reference to the manual flame sterilization procedure, namely 5, 10, or 20 s, and 15, 25, or 35°, respectively. All experiments were performed at room temperature in four replicates, and the average values were calculated.

### Measurement of CO2 concentration generated by flame sterilization procedure mimicked using CDMSS and AAFS

In order to investigate the effect of only the flame sterilization operation on the shake-flask culture without the other sampling processes, we connected a 99.8% (v/v) CO2 gas cylinder to the AAFS to aerate intermittently the headspace of the shaking flask15. The CO2 concentration in the gas phase of the flask was monitored using CDMSS under various intermittent aeration conditions, and the maximum CO2 concentration was recorded. All experiments were performed at room temperature in four replicates, and the average values were calculated.

### Computational domain and settings

Supplementary Fig. S6 online shows the schematics of the computational domain. A Bunsen burner, with a pipe size of 12.7 mm for its outer diameter and a thickness of 1.00 mm, was installed upwards. The capacity of the Erlenmeyer flask was 500 mL, and it was empty. The lower part of the flask tip was installed 60 mm above the burner rim top and the flask was tilted 60 °. The X, Y, and Z directional sizes of the computational domain were 450 mm, 320 mm, and 370 mm, respectively, and the domain was divided into 920,000 cells.

Calculations were conducted by using Fluent 2019R1 (ANSYS Inc., Canonsburg, U.S.A.). The physical modeling was as follows. The transport equations of mass, momentum, energy, and species were solved as laminar flow, and CH4, O2, CO, CO2, H2O, and N2 were considered chemical species, with their thermal properties depending on temperature. Westbrook and Dryer’s two-step reaction mechanism was employed32. Gravity was also considered. The time step was 5 × 10−4 s and the maximum iteration number per time step was 20. The calculation was continued until the flow time reached 20 s.

As boundary conditions, the inlet flow rate of the Bunsen burner was 6.53 × 10−5 m3/s at standard state. The fuel was CH4 and the oxidizer was air, which consisted of 21.0% O2 and 79.0% N2. The equivalence ratio Φ of the premixed gas was 2.00. Here, the equivalence ratio is defined as $$\varPhi =\frac{\frac{{X}_{CH4}}{{X}_{O2}}}{{\left(\frac{{X}_{CH4}}{{X}_{O2}}\right)}_{st}}$$, where Xi is the mole fraction of species i, the subscript st denotes stoichiometric conditions, and $${(\frac{{X}_{CH4}}{{X}_{O2}})}_{st}$$ is 0.5. For the remaining lower interface, 0.100 m/s was set as the upward ambient co-flow; the component was air. The temperature of these gases was 26.85 °C. The surrounding boundaries other than the lower parts were set as pressure outlets, and when inflowing from outside, the gases were air at 26.85 °C. The flask and burner were made of borosilicate glass and 304 stainless steel, respectively. For the borosilicate glass, the density, heat capacity, and thermal conductivity were 2,190 kg/m3, 740 J/kg-K, and 1.38 W/m-K, respectively. The corresponding values for the stainless steel were 7,930 kg/m3, 590 J/kg-K, and 16.7 W/m-K, respectively.

In the simulation process, a premixed flame was first formed without installation of the flask. After the flask was positioned, the temperature of the flask and inside the flask was set as ambient temperature and the inside filled with air. This time was defined as t = 0 s and a transient calculation was conducted.

### Empirical verification

A simple experiment was also conducted for verifying the accuracy of the simulation. The sizes and positions of the burner and flask were the same as those in the simulation in the previous section. The fuel was CH4 with a purity of > 99.5% and the oxidizer was ambient air. The flow rate of the fuel was controlled by mass flow meter, which was 0.680 L/min at standard state.

Supplementary Fig. S7 online compares the flame shapes between CH4 distributions in the simulation and a direct photograph of the experiment, without the flask. The flame shape in the simulation qualitatively corresponds with that in the experiment. Two flames, an inner and an outer flame, are formed. The height of the inner flame is approximately 40 mm. For the outer flame, although the height cannot be accurately measured because its tip is unclear, it is presumed to be approximately 80 mm. Supplementary Fig. S8 online shows the experimental setup for heating the flask. The flask was suspended by thin wires. Four K-type sheathed thermocouples (1.0 mm in diameter) were installed inside the flask. In order to measure the surface temperature of the flask using a non-contact infrared thermography device, the flask was colored black. Supplementary Fig. S9 online shows the temperature along the center axis of the flask at t = 20 s. In the experiment, these temperatures were measured at heights of 42, 72, 102, and 142 mm (see Supplementary Fig. S9 online, closed red circle). The temperatures for both the simulation and experiment were ambient from 0–100 mm. At > 120 mm, the temperature drastically increased in the simulation, approaching 1,400 °C at 172 mm. In contrast, in the experimental result, the temperature at 142 mm was 75.2 °C, which was 130 °C lower than the corresponding simulation. There are two main reasons that the temperature in the simulation was higher. First, a two-step reaction mechanism was used. For Φ = 2.0, the adiabatic flame temperature was 270 °C higher than the value when employing detailed mechanisms. Second, the dilution of burned gas by co-flows was different from that occurring in a real situation. Here, 0.1 m/s was set as the co-flow velocity because a calm flow field was assumed; flow fields might be more turbulent in the laboratory. In addition, according to the infrared thermography device, the maximum temperature of the flask exceeded 500 °C at t = 180 s, but the area where the temperature was > 100 °C was limited to the neck regions with heights > 130 mm.

As indicated above, we confirmed that our simulation can qualitatively evaluate the behavior of burned gas around the flask.

### Microorganisms and medium

E. coli K12 IFO3301, P. saccharophila NBRC103037, A. pasteurianus NBRC3283, and S. cerevisiae IFO0309 were used in this study. The LB medium used for E. coli cultivation consisted of (g/L): tryptone, 10; yeast extract, 5; and NaCl, 5. The R2A medium used for P. saccharophila cultivation consisted of (g/L): yeast extract, 0.5; peptone, 0.5; casamino acids, 0.5; glucose, 0.5; soluble starch, 0.5; sodium pyruvate, 0.3; K2HPO4, 0.3; and MgSO4·7H2O, 0.05. The NBRC no. 804 medium used for A. pasteurianus cultivation consisted of (g/L): hipolypepton, 5; yeast extract, 5; D-glucose, 5; MgSO4·7H2O, 1. The YM medium used for S. cerevisiae cultivation consisted of (g/L): D-glucose, 10; peptone, 5; yeast extract, 3; and malt extract, 3.

### Inoculum preparation

A loop-full of E. coli IFO3301, P. saccharophila NBRC103037, A. pasteurianus NBRC3283 or S. cerevisiae IFO0309 slant culture was inoculated into a 500 mL Erlenmeyer flask containing 100 mL of LB, R2A, NBRC no. 804 or YM media, respectively. The samples were then cultivated at 30 °C on a rotary shaker with a 70-mm shaking diameter at 100 rpm for 6.75 h (E. coli) or 20 h (P. saccharophila) and 200 rpm for 72 h (A. pasteurianus) or 14 h (S. cerevisiae). Glycerol stocks were prepared by adding the culture medium to glycerol (final concentration 20% [v/v]) and stored at −80 °C.

### Culture conditions

One millilitre of each glycerol stock was inoculated into a 500 mL Erlenmeyer flask containing 100 mL of 4 media types and cultured at 30 °C on a rotary shaker with a 70-mm shaking diameter at 100 rpm (E. coli and P. saccharophila) or 200 rpm (A. pasteurianus and S. cerevisiae). For air permeability, the Erlenmeyer flask was equipped with a BIO-SILICO N-38 sponge plug (Shin-Etsu Polymer Co., Ltd, Tokyo, Japan; breathable culture-plug). After each sampling of the culture broth, 99.8% (v/v) CO2 gas was aseptically aerated into each flask headspace using the AAFS for either 90 s or 30 s at 30 mL/min.

### Sampling of culture broth during shake-flask culture

Conventional sampling methods, which include interruption of shaking, opening the culture-plug, and flame sterilization on a clean bench, are known to affect the cultured microorganisms. Therefore, sampling during the shake-flask culturing was performed with the system shown in Supplementary Fig. S10 online, which is similar to the CDMSS sampling method5. Specifically, the sampling unit and flask were autoclaved independently. Then, the sampling unit, flask, three-way plug, and 50 mL syringe with a 0.2 μm filter were assembled on a clean bench, as shown in Supplementary Fig. S10 online. The three-way plug, 50 mL syringe, and 0.2 μm filter were disposable. Sampling of the shake-flask culture was performed by attaching a syringe to a three-way plug connected to a sampling tube. After collecting the culture broth, in order to ensure its removal from the sampling tube, a volume of fresh air (0.2 μm filtered) was used to flush the broth out of the tube. The sampled culture broth was stored at −80 °C until used for analysis.

### Measurement of culture factors

The U.O.D.600 and pH of the culture broth, which was sampled without interruption of shaking, were measured using a V-570 spectrophotometer (JASCO, Tokyo, Japan) and a pH meter (HORIBA, Kyoto, Japan), respectively. In order to minimize the decrease in volume of the culture broth due to repeated sampling, the total sampled volume did not exceed 10% (v/v) of the total volume of the initial culture medium. All measurements were performed in duplicate. In the case of P. saccharophila, the culture broth contained cell aggregates; these aggregates were disrupted by vortex and ultrasonication for 10 min at 4 °C prior to absorbance measurements.

## References

1. 1.

Kluyver, A. J. & Perquin, L. H. C. Methodology in the research of mold metabolism. Biochemische zeitschrift 266, 68–81 (1933).

2. 2.

Shiota, H. & Sakaguchi, K. Studies on the production of mold protease in the submerged culture. Part I. Nippon Nogeikagaku Kaishi 23, 426–429 (1950).

3. 3.

Reynoso-Cereceda, G. I., Garcia-Cabrera, R. I., Valdez-Cruz, N. A. & Trujillo-Roldan, M. A. Shaken flasks by resonant acoustic mixing versus orbital mixing: Mass transfer coefficient kLa characterization and Escherichia coli cultures comparison. Biochem. Eng. J. 105, 379–390 (2016).

4. 4.

Takahashi, M. & Aoyagi, H. Practices of shake-flask culture and advances in monitoring CO2 and O2. Appl. Microbiol. Biotechnol. 102, 4279–4289 (2018).

5. 5.

Takahashi, M., Sawada, Y. & Aoyagi, H. Development of a circulation direct sampling and monitoring system for O2 and CO2 concentrations in the gas–liquid phases of shake-flask systems during microbial cell culture. AMB express 7, 163 (2017).

6. 6.

Scheidle, M., Klinger, J. & Buchs, J. Combination of on-line pH and oxygen transfer rate measurement in shake flasks by fiber optical technique and Respiration Activity Monitoring System (RAMOS). Sensors 7, 3472–3480 (2007).

7. 7.

Schneider, K., Schutz, V., John, G. T. & Heinzle, E. Optical device for parallel online measurement of dissolved oxygen and pH in shake flask cultures. Bioproc. Biosystems. Eng. 33, 541–547 (2010).

8. 8.

Flitsch, D., Ladner, T., Lukacs, M. & Buchs, J. Easy to use and reliable technique for online dissolved oxygen tension measurement in shake flasks using infrared fluorescent oxygen-sensitive nanoparticles. Microb. Cell Fact. 15, 45 (2016).

9. 9.

Bauer, I., Poggendorf, I., Spichinger, S., Spichinger-Keller, U. E. & John, G. T. Novel single-use sensors for online measurement of glucose. Bioprocess Int. 10, 56–60 (2012).

10. 10.

Anderlei, T. & Buchs, J. Device for sterile online measurement of the oxygen transfer rate in shaking flasks. Biochem. Eng. J. 7, 157–162 (2001).

11. 11.

Anderlei, T., Zang, W., Papaspyrou, M. & Buchs, J. Online respiration activity measurement (OTR, CTR, RQ) in shake flasks. Biochem. Eng. J. 17, 187–194 (2004).

12. 12.

Schulte, A. et al. Parallel online determination of ethylene release rate by Shaken Parsley cell cultures using a modified RAMOS device. BMC Plant. Biol. 18, 101 (2018).

13. 13.

Takahashi, M. & Aoyagi, H. Monitoring of CO2 and O2 concentrations in the headspace of Sakaguchi flasks during liquid culture of microorganism. Appl. Microbiol. Biotechnol. 102, 6637–6645 (2018).

14. 14.

Buchs, J. Introduction to advantages and problems of shaken cultures. Biochem. Eng. J. 7, 91–98 (2001).

15. 15.

Takahashi, M. & Aoyagi, H. Effect of intermittent opening of breathable culture plugs and aeration of headspace on the structure of microbial communities in shake-flask culture. J. Biosci. Bioeng. 126, 96–101 (2018).

16. 16.

Kozlov, V. E., Starik, A. M & Titova, N. S. Modeling study of pollutant formation in homogeneous methane and syngas fueled combuster, 29th Congress Of The International Council Of The Aeronautical Sciences (2014).

17. 17.

Blombach, B. & Takors, R. CO2 - Intrinsic Product, Essential Substrate, and Regulatory Trigger of Microbial and Mammalian Production Processes. Front. Bioeng. Biotechnol. 3, 108 (2015).

18. 18.

Losen, M., Frolich, B., Pohl, M. & Buchs, J. Effect of oxygen limitation and medium composition on Escherichia coli fermentation in shake-flask cultures. Biotechnol. Progr. 20, 1062–1068 (2004).

19. 19.

Jones, R. P. & Greenfield, P. E. Effect of carbon-dioxide on yeast growth and fermentation. Enzym. Microb. Technol. 4, 210–222 (1982).

20. 20.

Gamboa-Suasnavart, R. A. et al. The metabolic switch can be activated in a recombinant strain of Streptomyces lividans by a low oxygen transfer rate in shake flasks. Microb. Cell Fact. 17, 189 (2018).

21. 21.

Arul Jose, P., Sivakala, K. K. & Jebakumar, S. R. Formulation and Statistical Optimization of Culture Medium for Improved Production of Antimicrobial Compound by Streptomyces sp. JAJ06. Int. J. Microbiol. 2013, 526260 (2013).

22. 22.

Batista, K. A. & Fernandes, K. F. Development and optimization of a new culture media using extruded bean as nitrogen source. MethodsX 2, 154–158 (2015).

23. 23.

Anand, A. & Aoyagi, H. Estimation of microbial phosphate accumulation abilities. Sci. Rep. 9, 4879 (2019).

24. 24.

Tsuji, M. Cold-stress responses in the Antarctic basidiomycetous yeast Mrakia blollopis. R. Soc. Open sci. 3, 160106 (2016).

25. 25.

Ogbonna, C. N., Aoyagi, H. & Ogbonna, J. C. Isolation and identification of Talaromyces purpurogenus and preliminary studies on its pigment production potentials in solid state cultures. Afr. J. Biotechnol. 16, 672–682 (2017).

26. 26.

Endo, R. & Aoyagi, H. Adsorption preference for divalent metal ions by Lactobacillus casei JCM1134. Appl. Microbiol. Biotechnol. 102, 6155–6162 (2018).

27. 27.

Xue, W. J. et al. Production of biomass and recombinant human-like collagen in Escherichia coli processes with different CO2 pulses. Biotechnol. Lett. 31, 221–226 (2009).

28. 28.

Ude, C. et al. New perspectives in shake flask pH control using a 3D-printed control unit based on pH online measurement. Sensor Actuator B Chem. 221, 1035–1043 (2015).

29. 29.

Krause, M., Neubauer, A. & Neubauer, P. The fed-batch principle for the molecular biology lab: controlled nutrient diets in ready-made media improve production of recombinant proteins in Escherichia coli. Microb. Cell Fact. 15, 110 (2016).

30. 30.

Philip, P. et al. Systematic evaluation of characteristics of the membrane-based fed-batch shake flask. Microb. Cell Fact. 16, 122 (2017).

31. 31.

Philip, P. et al. Parallel substrate supply and pH stabilization for optimal screening of E. coli with the membrane-based fed-batch shake flask. Microb. Cell Fact. 17, 69 (2018).

32. 32.

Westbrook, C. K. & Dryer, F. L. Simplified reaction mechanisms for the oxidation of hydrocarbon fuels in flames. Combust. Sci. Tech. 27, 31–43 (1981).

## Acknowledgements

This work was supported by Noda Institute for Scientific Research Young Investigator Research Grant and JSPS KAKENHI Grant-in-Aid for Early-Career Scientists (20K15097) (grants to MT). This work was also supported in part by JSPS KAKENHI Grant-in-Aid for Scientific Research B (19H03086), the Sumitomo Electric Industries Group Corporate Social Responsibility Foundation, and Noda Institute for Scientific Research (grants to HA). The manuscript has been edited carefully by native-English-speaking professional editors from Editage, a division of Cactus Communications. We would like to thank Editage (www.editage.jp) for English language editing.

## Author information

Authors

### Contributions

M.T. and H.A. designed the research. H.A. supervised the research. M.T. designed the wet experiments and performed all wet experiments without Computational Fluid Dynamics. T.H. designed and performed the dry experiments. M.T., T.H., R.T. and H.A. analyzed the data. M.T., T.H., R.T. and H.A. wrote the manuscript. All authors read and approved the final manuscript.

### Corresponding author

Correspondence to Hideki Aoyagi.

## Ethics declarations

### Competing interests

The authors declare no competing interests.

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

## Rights and permissions

Reprints and Permissions

Takahashi, M., Honzawa, T., Tominaga, R. et al. Analysis of the influence of flame sterilization included in sampling operations on shake-flask cultures of microorganisms. Sci Rep 10, 10385 (2020). https://doi.org/10.1038/s41598-020-66810-3