Effects of Oxygen Concentration on Glucose Sensor Response

Blood glucose monitoring is important both for the patients and the health care providers in order to keep track of blood glucose levels and to apply necessary medication and treatment accordingly.   One common method to determine blood glucose is with the glucose oxidase (GOx) enzyme sensor.   In presence of the enzyme and a cofactor which is naturally oxygen, the enzyme's oxidation of β-D glucose produces hydrogen peroxide which is oxidized at an electrode surface giving up two electrodes. Current from the electrons becomes the analytical signal, and it is proportional to the concentration of hydrogen peroxide produced during the reactions.  The assumption is that oxygen is in excess and the production of peroxide is glucose limited, so that the peroxide oxidation current is also directly proportional to the glucose concentration. The glucose/oxygen ratio must be kept less than 1 to keep the peroxide producing reaction rate glucose-limited.  If the ratio becomes more than 1, the reaction rate becomes oxygen-limited, and the relationship between the signal and the glucose concentration is lost. One solution has been to eliminate the dependence on oxygen by replacing oxygen with an artificial mediator, and for in vitro use such as the glucose test strip, this approach has been successful.  Unfortunately, these mediators can be toxic, and their leaching out makes them unsuitable for in vivo use.  Another scheme that uses oxygen can be made to work by lowering sensitivity of the probe.  This reduces the enzyme's demand for oxygen, but it fails to detect the glucose accurately.  We propose an alternate solution in which an electrode at micro-range proximity to the enzyme electrode generates oxygen by electrolyzing water.  This artificially produced oxygen relieves enzyme's dependence on endogenous oxygen.

For the scheme to work, GOx must be localized onto a single electrode in an electrode array.  Immobilizing techniques such as dip-coating are not suitable leaving electropolymerization as the best alternative.  Initially several attempts involved GOx entrapment into polypyrrole by electropolymerization on a Pt electrode.  These attempts were unsuccessful in detecting glucose, and colorimetric enzyme activity tests confirmed low activity of enzyme in the film. Possible reasons of failure are too little GOx being entrapped or peroxide being unable to approach the electrode.  To determine the effects of oxygen on the GOx sensor, we temporarily abandoned electropolymerization and used dip coating.  A glutaraldehyde/bovine serum albumin/GOx film was successful in detecting glucose.  Figure 1 shows an oxygen-stress experiment using a brain-level glucose concentration of 0.5mM. After glucose signal stabilizes, the glucose-solution is completely deaerated by nitrogen purging.  As the oxygen decreases, the glucose/oxygen ratio becomes >1 , and the GOx reaction rate becomes oxygen-limited.  What is seen is the anodic current gradually diminishing as if the glucose concentration was decreasing.  At the point where the air starts dissolving back into the solution, the anodic current signal shows a gradual increase with increasing oxygen. The experiment was performed in the brain oxygen range between 0.05-3.0 ppm.  Had the oxygen been allowed to increase further, the original glucose signal would have been restored. 

Entrapment of GOx into poly-o-phenelenediamine was achieved by electropolymerization on Pt electrode which successfully detected glucose. First, the bare Pt electrode was polished using 1µm diamond polishing solution and was electrochemically pre-treated in 0.5 M H₂SO₄  with potential cycling between -0.21V and +1.19V vs. standard calomel electrode (SCE). The electropolymerizing solution was 5mM oPD and 500U/mL GOx in acetate buffer (pH 5.2, I= 0.2M).  Film was grown for 15 minutes at a constant potential of +0.65V vs. SCE without stirring. The electropolymerized enzyme electrode was washed in stirring 10mM Phosphate buffer solution (PBS) (pH 7.4) for 3 minutes and stored in PBS at +4^oC when not in use.

80ml of 10mM PBS was used as the electrolyte for the glucose-response test.  A 0.8mM glucose stock solution in 10mM PBS was prepared and kept overnight at room temperature before use.  A +0.7V potential was applied, and after an initial 5 minutes of signal recording at 0mM glucose, the solution was spiked every 5 minutes with required amount of glucose stock solution to advance the glucose concentration in the solution in 1mM increment.  As the glucose in the solution increases, more peroxide is produced which in turn produces a greater oxidation current.  Figure 2 shows the increase in signal upon glucose additions in the range of 0 to 7mM, a typical range for blood-glucose.  The data in figure 2 was used to construct the calibration curve shown in figure 3, which was linear with a sensitivity of 0.2119 µAmp/mM.