Wet Lab
1. PQQGDH Production:
pTRC99a vector containing the pyrroloquinoline quinone glucose dehydrogenase (PQQGDH) gene was transformed into Escherichia coli BL21 DE3 cells. A mixture of colonies was incubated into 3mL LB Media (Amp 100µg/mL) for 24hrs at 37℃. One percent of preculture was inoculated into 4, 100mL Flasks of LB Media. Large cultures were cultivated at 37 ℃ until the OD600~0.5 at which 200mL IPTG was added. Protein expression continued at 30 ℃ overnight. Cultures were centrifuged and the wet cells were resuspended in buffer (10mM potassium phosphate buffer;PPB pH 7.0) and lysed twice via French Pressure (Glen Mills) at 2000 psi for 30 seconds. Following centrifugation, ultracentrifugation, and dialysis overnight (10mM PPB pH 7.0) cell lysate was purified via cation exchange chromatography (Resource S 1mL, Elution Buffer 10mM PPB. 500mM NaCl, pH 7.0).
2. Electrochemical Analysis:
Electrochemical cleaning:
Bare gold disc electrodes were polished (0.05 µm alumina powder), chemically washed (50 mM KOH Solution (1:5 300mM KOH:H2O2) soak for 45 Min), and electrochemically washed (Cyclic Voltammetry (CV); -1.2 to -0.2 V (50 V/s; 15 cycles) in 50 mM KOH.
Enzyme immobilization:
3 μL pf glucose dehydrogenase (GDH) mixed with 22 μL Ketjen black (KB) was prepared for the anode while multicopper oxidase (mCOP) mixed with KB at the same volumes was prepared for the cathode. Clean gold disc electrodes were coated three times with 7 μL of the corresponding enzyme/KB mixture with 45-minute drying periods between each coat. This was followed by 1-hour glutaraldehyde deposition and 20-minute TRIS soak.
Ferricyanide Power Test - Version 1:
Pairs of PQQGDH immobilized anode and MCO immobilized cathode were placed in electrochemical cells with 20mM glucose and 100 mM PPB. The electrochemical cell was connected to a resistor box at 1MΩ resistance. Open-circuit potentiometry (OCP) measured the drop in potential between the anode and cathode (mV) over time. This experiment was repeated for concentrations of ferricyanide at 0 μM, 25μM, 50μM, 75μM, and 100 μM.
PQQGDH anode and mCOP cathode demo
Pyocyanin Power Test - Version 2:
Pairs of PQQGDH immobilized anode and MCO immobilized cathode were placed in electrochemical cells with 100 mM PPB. Open-circuit potential (OCP) was measured by the decrease in the potential between the anode and cathode (mV) over time. Pyocyanin was added to the buffer solution at 0 μM, 25μM, 50μM, 75μM, and 100 μM and tested without glucose. Then, each concentration of pyocyanin was tested with 20 mM glucose in the solution. The electrochemical cell was connected to a resistor box set at 1 kΩ and then raised to 100 kΩ and 1 MΩ.This produced OCP data for concentration of pyocyanin 0-100 μM both with and without glucose and at three different resistances.
3. Activity Testing
Clean bare gold disc electrodes were placed into a solution of 10 mM PPB with 1mM Ca2+ ion and 100 mM Cl- ion supplement. Cyclic voltammetry was run -0.5 to 0.3 at 50 mV/s for 2 cycles. Argon gas was released into the solution to pump oxygen out of the system. Pyocyanin at 100 μM concentration, PQQ at 1 μM concentration, and the two reagents combined were tested.
4. Hydrogel Electrodeposition
Hydrogel solution was prepared with 1% w/v chitosan and 1% v/v acetic acid. Clean gold disc electrodes were placed into the hydrogel solution. Chronoamperometry was run for 7 minutes to electrodeposit the hydrogel onto the electrodes.
Dry Lab
Data sheets from the pyocyanin and PQQGDH power test were k-clustered by voltage, where k=6, in order to allow for the data to naturally separate into groups corresponding to their resistance level. K was chosen greater than 3 although there were 3 resistances to account for noise and outliers in the data, even among clusters. Each graph was then visually expected to decide which cluster medians should be counted into the mean of interest. For example, if the 10 ohms time section had 3 clusters, the mean of the 3 medians of the 3 clusters would be considered as the central tendency of the 10 ohms section. In other words, we visually checked the graphs, grouped similar clusters together based on when resistance was changed, and then found the average of their medians to represent the central tendency for each section. When k=6 K-means could not adequately cluster the data, the data was manually clustered into other k values until an adequate value was found—one that could best represent a series of times while including as few outliers as possible. Medians, variances, and sample sizes of each cluster of interest were recorded. Next, these data were used to create graphs and run statistical tests, principally least-squares linear regression, ANOVA, and post hoc Turkey HSD tests.
Recommendations for future data collection:
Mean Voltage of Every Trial in the format:
GLC or NGLC signifies glucose or no glucose,
25, 50, 75, or 100 signifies the pyocyanin concentration,
R1, R2, or R3 signifies the resistance level (low, medium, and high respectively),
E1, E2, E3, and C, signify electrode number or control
Linear Regression was done to determine if a trend exists between the voltage and any of the variables. Control data was not considered in regression so as to not alter the results. There was a significant relationship between resistance and mean voltage (p < 2.139e-05, R2=0.269) but not a significant relationship between the concentration of pyocyanin and mean voltage (p>.68). This suggests that pyocyanin concentration did not appreciably alter the voltage differences.
ANOVA was done to compare means between concentrations of pyocyanin, resistance levels, and electrodes, and whether glucose was present or not. Control data was not included so as to not influence the data except when doing ANOVA on electrode number. There was no difference in means between glucose and no glucose (p> .718). There was also no difference in mean voltage between pyocyanin concentrations. Along with the linear regression results, this further suggests pyocyanin concentration didn't significantly influence mean voltage differences. There was no difference in means between control and electrodes 1 and 2. However, Electrode 3 was significantly different than both the control and the other 2 electrodes (p < 0.0000778). This informs us the experimental conditions had no effect on the mean voltage when measured with those two electrodes, but there was a difference when measured with electrode 3. Electrode 3 was significantly different than electrodes 1 and 2, suggesting there is a bias in an electrode, as the electrodes were not designed to measure the voltage differently.