Background

Measuring Glucose with a Liquid Handling Station - Implementation of a Workflow

This project implements a glucose measurement workflow in a Hamilton Liquid-Handling-Station.

It includes a small introduction into liquid handling robots and what they are used for. It describes why they are a powerful tool to conduct high throughput experiments and further summarizes some challenges that must be tackled when implementing an automated workflow.

We further describe the working principle of the glucose sensor and the experimental structure that was used to gather training data for a model to predict glucose concentrations from samples within a short measurement time. The model is described elsewhere (insert hyperlink here). In the Materials and Methods section we describe the steps to achieve the automated measurement of glucose samples and describe the workflow in detail. This allows the reader to implement and understand the current state of this workflow and to further work on it.

Literature Review

Glucose Oxidase (GOx)

The enzyme glucose oxidase (GOx) is an oxygen dependent oxidoreductase for β-D-Glucose (EC 1.1.3.4.). GOx was first discovered by D. Müller in 1928 (Müller, 1928) in extracts from Aspergillus niger. In addition to fungi such as Aspergillus niger, various Penicilium spp., Taloromyces flavus and others, GOx is naturally produced by some insects such as the honeybee and grasshopper (Chun Ming et.al., 2008). The main function of Gox in nature is the production of hydrogen peroxide as anti-bacterial and anti-fungal agent. GOx is a dimeric glycoprotein with a molecular weight around 160kDa depending on the level of glycosylation. It consists of two identical subunits, each of the subunits bounds non-covalently with a flavine adenine dinucleotide (FAD) (Figure 1). This redox active co-enzyme acts as an electron carrier during the catalysis.

Figure 1: X-ray crystal structure analysis of the homodimer of the glycoprotein glucose oxidase (GOx) bound to the co-enzyme FAD (Bartletta et al, 2017)

GOx catalyses the reaction of ß-D-glucose and oxygen to D-glucono-1,5-lactone, and hydrogen peroxide as shown in figure 2. The first step of this redox reaction is the oxidization of glucose to D-glucono-1,5-lactone, while the co-factor FAD is reduced to FADH2. This is immediately followed by an electron transfer from FADH2 to the electron acceptor oxygen to release the hydrogen peroxide as product and restores the FAD.

GOx is highly selective to ß-D-glucose, while the activity with most other substrates results in reaction rates lower than 2% of ß-D-glucose. This specificity is attributed to the structure of the enzyme. The catalytic centre with the FAD enzyme is located in a narrow passage between the two homodimers and prevents a nonspecific binding with other redox species (Bartletta et al, 2017). The anomer α-D-glucose has an approx. 150-times lower reaction rate and the second-best rate with only 25 % of ß-D-glucose shows 2-desoxy-D-glucose (ß-D-glucose with a H instead of OH at the C-2). (Raba and Mottola, 2006)

Figure 2:  GOx catalysed reaction of ß-D-glucose and oxygen to D-glucono-1,5-lactone and hydrogen peroxide.

In addition to the high substrate specificity, the GOx shows a rapid turnover rate and high stability to external conditions. The enzyme shows the best function at pH 5, but it is stable without loss of function in a range between pH 2 to pH 8. Its unstable at temperatures above 40°C but it can be protected against thermal denaturation by polyhydric alcohols. (Ye et al.,1988; Nakamura et al., 1976)

There are different modern medical and industrial applications for GOx. In addition to the use GOx in the glucose measurement with biosensors, further fields of applications are the wine production (decreased alcohol formation by metabolizing glucose), the bread and baking industry (to improve dough structure by hydrogen peroxide formation), the medical regulation of the blood glucose level in combination with insulin and for food preservation (Bauer et al 2022, Röcker et al 2016, El-Rashidy et al 2015). The most important hosts for the industrial production of GOx are the species Aspergillus spp. and Penicillium spp., such as A. niger and P. amagasakiense.

GOx Glucose Sensors

The measurement of glucose in biosensors is based on the interaction with one of the three enzymes, hexokinase, glucose-1-dehydrogenase and GOx. The enzymes differ in their turnover rate, redox potential, selectivity for glucose, cofactors, costs and resistance to environmental influences. However, GOx became the standard enzyme for most of the biosensors, due to its low cost, high resistance to high/low pH Values and temperature and the high specificity for ß-D-glucose. The majority of the glucose biosensors are electrochemical sensors, which can be subdivided into amperometric, potentiometric or conductometric sensor types. Amperometric sensors measure glucose concentration by applying a voltage (0,8 V with H2O2; <0,8 V with mediators) between a reference electrode and the measuring electrode. The voltage leads at the measuring electrode to the cleavage of hydrogen peroxide to hydrogen, oxygen and electrons. The resulting current flow is proportional to the glucose concentration and can be detected by an amperemeter (Liao et al, 2007). Potentiometric biosensors…

Furthermore, the sensors can be divided into three different generations which are schematically represented in figure 3.

In the first generation of GOx based sensors, the dissolved, molecular oxygen acts as electron acceptor. The enzyme catalyses the “natural” reaction described above and the reduced FADH2 forms, with molecular oxygen, hydrogen peroxide.

In the second generation, a mediator takes over the role of oxygen and acts as an electron acceptor. The advantage of a mediator is that the oxygen concentration of the medium no longer has an influence on the measurement. A variety of electron mediators are known, including ferrocene, ferricyanide, quines, methylene blue, thionine and tetracyanoquinodimethane (Yoo et al, 2010, Suzuki et al, 2020). However, the mediator must be present in close proximity to the electrode and the enzyme, which can be achieved by an insoluble mediator that is immobilized on the sensor surface (figure 3, 2.5th Generation). Further strategies to improve the electron transport are enzyme wiring by electron conducting redox hydrogels and nanotubes (Wang et al, 2012). The third generation does not require an electron acceptor at all. The electrons are transferred to the electrode by direct electron transfer.

Figure 3: Generation of GOx based Biosensors (Suzuki et al 2020)

On Liquid Handling Robots and why they are used

Liquid handling robots are used to transfer, mix, and store samples, reagents, and other types of liquid in an automated manner. In life science related experiments precision of liquid handling is key to successful experiments and tricky when done by hand. Human errors can be avoided when using a programmed robot. (He et al, 2016, Silva et al, 2021) Sample material and volume are often limited, and reagents can be expensive. When working with small volumes, pipetting is prone to errors when done by hand. Small errors can lead to significant in-batch and batch-to-batch variation. The liquid handling robot performs well on the aspiration and dispersion of small volumes of various liquids. This also leads to low sample volume consumption. The volumes that are pipetted can be precisely controlled to achieve good quality of data and accurate results when analysing compounds and samples with sensitive assays and sensors. (Kong et al, 2012, Silva et al, 2021) Another great advantage of liquid handling robots is that they enable automation of time-consuming workflows. Pipetting steps can be parallelized using the multiple channels of the robot and multiwell plates as containers. Thus, the robot can handle large numbers of samples in a short period of time. The less time a scientist spends on assay workflows in the lab, the more time can be spent on data processing and the design of new experiments. Therefore, it is attractive to automate labour-intensive sample processing steps such as diluting. (Kong et al, 2012, Silva et al, 2021, Sakar and Kumar, 2021)

In summary, automated workflows performed by liquid handling robots reduce time, cost and errors and ensure good quality of data, reproducibility and standardization of various experimental operating and handling procedures. High throughput experiments can be conducted to gather a large quantity of reliable information and data in a short period of time. (Hemmerich et al, 2018, Silva et al, 2016)