Results and Discussion

Glucose concentation measurement:

Figure 8 shows the glucose concentration of the 18 standards determined with the CEDEX Bio HT (Cobas Integra 400). The graph shows a deviation between the measured concentration and the target value. This can be explained by the fact that the use of glucose monohydrate was not taken into account when calculating the glucose standards. Therefore, the expected value was added to the chart, which takes into consideration the 10 % deviation due to the use of glucose monohydrate (M(Glucose)= 180.156 g/mol, M(H20)= 18.01528 g/mol) instead of glucose.The measured glucose concentration shows a very good agreement with the expected value, so the dilution series worked well. The deviation from the target value due to the above-mentioned error is negligible for the tests, because the glucose values measured with the CEDEX are stored as reference values in the model and this does not lead to any impairment.

Figure 8: Measured glucose concentration with the Cedex Bio HT (COBAS INTEGRA 400). In addition to the measured concentration (red) and the target concentration (black), the expected concentration (green) is also plotted. The expected value is slightly below the target value because glucose monohydrate was used for weighing, but this was forgotten in the calculations

During the development and implementation of the above presented workflow, the method was continuously tested for functionality. Thereby, challenges and hurdles were identified quickly. One of the first challenges faced was that the sensor chamber that was 3D-printed beforehand was leaking. Thereby, two problems occurred. First, storing the sensor in buffer solution after usage (see Ending the Workflow) was not possible due to buffer leaking out of the camber. This was possibly the reason for the first sensor to be damaged since after storage overnight, it did not record a signal after reinitialization of the test runs. Secondly, leakage can cause the cross-contamination between sample runs as buffer (3 g/L glucose) and sample volume can remain in so called dead zones. This could result in altered concentrations, resulting in lower data quality of the set of reference standards used to train the model. To avoid leakage, a new sensor chamber design was developed.

A second challenge occured when handling small sample volumes in open containers over a longer time period as planned for this experiment (1 week). Evaporation was observed in recognizable amounts during the Evavoration Test. To identify the effect of evaporation on the quality of data, a simplistic model was used to calculate the change in concentration of glucose over time. It was assumed that glucose does not crystallize but remains in solution which would have the largest effect on the concentration. The evaporation rate was determined using water as a reference fluid. Effects of sample volume, as well as glucose concentration on the evaporation rate was neglected in the model. Figure 9 shows the hypothetical change in concentration and volume over a duration of one week (168 hrs). Glucose concentration could increase by about 32 %, while the volume decreased by around 31 %. Again, this is a very simplistic model and only gives an idea of the potential outcome if not considering evaporation when using the data obtained.

Figure 9: Exemplary calculated time course of the glucose concentration (black) and sample volume (red) of the 1 g/L glucose standard in the Hamilton Liquid Handling Station determined by the evaporation rate of water.

However, if the rate of evaporation can be precisely evaluated, the change in concentration due to evaporation could be used to increase the sample concentrations that are measured without increasing the sample number. Though, it requires precise modelling of evaporation which exceeds the extend of this study. Another strategy which could help reducing the evaporation effect would be to change the type of labware that holds the reference standards to a 96-well plate. These are available with lids that are only opened, when sample is removed for measurements, so that evaporation during measurements is reduced by the enclosed environment. Other factors such as cooling of the Hamilton Microlab Star would require excessive investments and are most likely not economical. Also, one workspace of the Hamilton robot is not enclosed completely so that heat exchange between the laboratory and the inside of the Hamilton takes place.

Outlook

During the internship, a workflow to measure the glucose concentrations of a set of reference standards with the Microlab Star Hamilton Liquid Handling Station was implemented. These measurements are to be used for generating a large dataset of glucose concentration measurements with the CITSens MeMo USB. It aims to minimize the hands on time in the lab and generate reliable, accurate and large datasets. The measured glucose concentrations will serve as a training dataset and a test dataset to implement a model which predicts the glucose concentration of a measured sample before the signal reaches equilibrium. In brief, the predicted end point value is obtained by the slope of the signal over the measurement duration as the measurement takes place. The change in signal intensity is large within the first seconds to a minutes after the recording started but asymptotically approaches equilibrium, taking up to 10 minutes reaching it depending on glucose concentration.

Model-predicted endpoint concentrations could allow decreasing the sensor measurement duration, as one does not have to wait until the sensor reached equilibrium but gets a result after the first few minutes of measurement. This is especially important when considering the future application of the sensor serving as an at-line monitoring tool for the glucose concentration inside (mini)-bio-reactors used in the context of this research group. As glucose concentrations are a key parameter to monitor and control, a quick and accurate measurement is important. The next step after implementing and testing the developed workflow it is now time to generate a dataset which is used for the model development.

References

Bartlett, Philip N.; Al-Lolage, Firas A. (2018): There is no evidence to support literature claims of   direct electron transfer (DET) for native glucose oxidase (GOx) at carbon nanotubes or   graphene. In: Journal of Electroanalytical Chemistry 819, S. 26–37. DOI:   10.1016/j.jelechem.2017.06.021.

Bauer, Jacob A.; Zámocká, Monika; Majtán, Juraj; Bauerová-Hlinková, Vladena (2022): Glucose   Oxidase, an Enzyme “Ferrari”: Its Structure, Function, Production and Properties in the Light   of Various Industrial and Biotechnological Applications. In: Biomolecules 12 (3), S. 472.   DOI:10.3390/biom12030472.

El-Rashidy, L-A.; Bahlol, H. E. M.; El-Desoky, A. A. (2015): Improving Quality of Pan Bread by Using   Glucose Oxidase and Lipase Enzymes (5). In: Middle East Journal of Applied Sciences 2015 (5)  Pages: 1035-1043

He, Nongyue; Liu, Ting; Liu, Bin (2016): Technologies and Applications in Micro-Volume Liquid   Handling. In: Journal of nanoscience and nanotechnology 16 (1), S. 58–66.   DOI: 10.1166/jnn.2016.11681.

Kong, Fanwei; Yuan, Liang; Zheng, Yuan F.; Chen, Weidong (2012): Automatic Liquid Handling for Life   Science: A Critical Review of the Current State of the Art. In: SLAS Technology 17 (3),   S. 169–185. DOI: 10.1177/2211068211435302.

Liao, Cheng-Wei; Chou, Jung-Chuan; Sun, Tai-Ping; Hsiung, Shen-Kan; Hsieh, Jui-Hsiang (2007):   Preliminary investigations on a glucose biosensor based on the potentiometric principle. In:   Sensors and Actuators B: Chemical 123 (2), S. 720–726. DOI: 10.1016/j.snb.2006.10.006.

Müller D (1928) Oxidation von Glukose mit Extrakten aus Aspegillus niger. Biochem Z 199:136–170

Nakamura, S.; Hayashi, S.; Koga, K. (1976): Effect of periodate oxidation on the structure and   properties of glucose oxidase. In: Biochimica et biophysica acta 445 (2), S. 294–308. DOI:   10.1016/0005-2744(76)90084-x.

Raba, Julio; Mottola, Horacio A. (1995): Glucose Oxidase as an Analytical Reagent. In: Critical Reviews  in Analytical Chemistry 25 (1), S. 1–42. DOI: 10.1080/10408349508050556.

Röcker, Jessica; Schmitt, Matthias; Pasch, Ludwig; Ebert, Kristin; Grossmann, Manfred (2016): The   use of glucose oxidase and catalase for the enzymatic reduction of the potential ethanol   content in wine. In: Food Chemistry 210, S. 660–670. DOI: 10.1016/j.foodchem.2016.04.093.

Sarkar, Joyita; Kumar, Ashok (2021): Recent Advances in Biomaterial-Based High-Throughput   Platforms. In: Biotechnology journal 16 (2), e2000288. DOI: 10.1002/biot.202000288.

Silva, Tiago Castanheira; Eppink, Michel; Ottens, Marcel (2022): Automation and miniaturization:   enabling tools for fast, high‐throughput process development in integrated continuous   biomanufacturing. In: J of Chemical Tech & Biotech 97 (9), S. 2365–2375.   DOI: 10.1002/jctb.6792.

Suzuki, Nanami; Lee, Jinhee; Loew, Noya; Takahashi-Inose, Yuka; Okuda-Shimazaki, Junko; Kojima,   Katsuhiro et al. (2020): Engineered Glucose Oxidase Capable of Quasi-Direct Electron   Transfer after a Quick-and-Easy Modification with a Mediator. In: International journal of  molecular sciences 21 (3). DOI: 10.3390/ijms21031137.

Wang, Lisha; Gao, Xia; Jin, Lingyan; Wu, Qi; Chen, Zhichun; Lin, Xianfu (2013): Amperometric glucose   biosensor based on silver nanowires and glucose oxidase. In: Sensors and Actuators B:  Chemical 176, S. 9–14. DOI: 10.1016/j.snb.2012.08.077.

Wong, Chun Ming; Wong, Kwun Hei; Chen, Xiao Dong (2008): Glucose oxidase: natural occurrence,   function, properties and industrial applications. In: Applied microbiology and biotechnology  78 (6), S. 927–938. DOI: 10.1007/s00253-008-1407-4.

Ye, W. N.; Combes, D.; Monsan, P. (1988): Influence of additives on the thermostability of glucose   oxidase. In: Enzyme and Microbial Technology 10 (8), S. 498–502.   DOI: 10.1016/0141-0229(88)90028-2.

Yoo, Eun-Hyung; Lee, Soo-Youn (2010): Glucose biosensors: an overview of use in clinical practice. In:   Sensors (Basel, Switzerland) 10 (5), S. 4558–4576. DOI: 10.3390/s100504558.