We planned to design a reactor system that is suitable to investigate the electrochemical behaviour in bioprocesses. That includes the possibility to characterize mediators and different electrode materials on the one hand and the electron uptake into the cells on the other.
During our research we discovered the H-cell reactor that seemed to meet with our needs. (Park et al., 1999)
The advantages of an electrochemical measuring cell with separated compartments are that there is no mixing within the nascent products of electrolysis of the anode- and cathode compartment and the possibility to use different buffers in both compartments.
We approached two different concepts to realize the reactor construction. One of our H-cell reactors was constructed with the possibilities given to us by the facillities of our university. We instructed the glass workshop to modify two glass bottles by adding a glass-flange. Besides that the technical workshop build the lids from stainless steel. This approach had the advantage that we could influence the design and had to make precise design drawings especially for the connections in the lids.
The second H-cell reactor was a commercially available system by Adams & Chittenden scientific glass. The commercial system had a smaller volume and the benefit of a larger flange diameter. The necessary lids for that system were also custom design by our workshop. In figure 1 you can see both reactors in comparison.

Figure 1: Single parts of our self-designed H-cell reactors: 1 Custom designed lids that provide connections for: a pO₂-electrode, a pH-electrode, an entrance for reference and working electode, air output, heating coils and acid/ base input for pH control, 2 Heating coils, 3 Clamps for the flange connection 4 Sealing rings.

The H-cell is suitable for experiments concerning the investigation of mediator redox-characteristics and indirect electron transfer into electrotrophes.
In addition to the H-cell design we thought of an alternative reactor design that meets with the requirements of direct electron transfer. To enable direct electron transfer it is necessary that there is a large electrode surface provided to the microorganisms. Furthermore substrate limitation should be avoided. To meet with these requirements it is favourable to have an reactor that can be continiously driven. Our proposed solution is a flow cell reactor (FCR) which could be driven continiously.

Testing the set up

Our first experiments were carried out with a constant power supply and we measured the voltage input and the current. The set up is shown in figure 2.

Figure 2: Set up of our first experiments with the H-cell: 1 Ammeter 2 Power supply and voltmeter 3 Cathode compartment 4 Anode compartment

During our first experiments we filled both, the cathode- and the anode-space with phosphate buffer where neutral red was added to final concentration of 100 µM.
It turned out that we could not use the pH-electrode and the pO₂-electrode during our cultivations, because they affected the measurement. Especially the pO₂-electrode was not suitable in this set up, due to the fact that it is completely made of steel. It turned out that the electrode achieved a grounding of the system which set the lid under electric power. This resulted in a couple of unwanted oxidation processes at the weldseam of the lid. The consequence was to remove both electrodes from the system.
After the customization of the system we carried out a few test runs with differetn electrode materials. At this stage of the project we optimized the attachment of the electrodes and the isolation of the wire which was layed within a silicon tube through the lid.
THe first experiments showed that neutral red is reducable within out set up. The problems that occure if you work with a constant power supply are that the cell potential can not be kept at the same level during the experiments. The dynamic of the electrochemical reactions introduce unwelcome variabilities that cause fluctuations in the potential. Especially the presence of proliferating microorganisms can enhance this effect.

Different electrode materials

We tested different electrode materials for their potential to work in our reactor. We decided to investigate fabric carbon, fabric fleece and platinum electrodes. The different materials are shown in figure 3.

Figure 3: Different electrode materials for characterization experiments: 1 Carbon fabric 2 Carbon fleece 3 Platinum electrode.

Carbonic materials have the advantage that they are relativly cheap and are available in huge amounts. The nature of the processing of the material has a major influence on its electrochemical behavior.
Carbon fabric is made up of individual fibres and has therefore a good stability. Another advantage is that the fibres can overcome quite a long distance due to the fact that they are made of one piece. This assures a good electrical conductivity.
The carbon fleece instead is thicker and provides a larger surface for the microorganisms to attach to the electrode material. This advantage goes at expense of stability and conductivity. The fleece is made of lots of single fibres which leads to a bad connection between them and therefore causes an unfavourable conductivity.

Cultivation - constant voltage

The first experiments in the H-cell reactor were performed under constant direct voltage. These experiments were carried out to test the set up with microorganisms. We investigated if E. coli was able to grow within the needed voltage range and if the different mediators influence the cells if a small electric current is applied.

Figure 4: Comparisson of the growth compatibility of the E. coli KRX WT strain when a voltage is applied or not. Both cultivations were performed in the H-cell reactor in M9 minimal media- xylose (50 mM). One of the cultivatons was performed with an applied voltage of -330 mV the other one was currentless. The optical density and the Xylose concentration were measured with technical duplicates.

The cultivation curves in figure 4 show that E. coli is barely influenced in its growth by an applied current. The optical density reaches the same final value and the Xylose consumption is identical.
These results lead to the conclusion that E. coli is not affected in growth by an applied voltage.
A difference can be observed in the growth between the E. coli wildtype and the constructed E. coli ΔdcuB::oprF when both strains are cultivated with a constant voltage of -330 mV. Both strains were cultivated in the H-cell reactor in M9 minimal media- xylose (50 mM) that was supplemented with neutral red to a final concentration of 100 µM.

Figure 5: Comparison of the growth of the E. coli KRX WT strain and the constructed E. coli ΔdcuB::oprF. Both strains were cultivated with a constant voltage of -330 mV in the H-cell reactor in M9 minimal media- xylose (50 mM) with 100 µM neutral red added. The optical density and the Xylose concentration were measured with technical duplicates.

The growth curves imply that the E. coli ΔdcuB::oprF strain finds advantageous conditions for its growth in comparison to the E. coli KRX wild type.
Maybe this effect occures due to the added mediator. To give a valid statement on this effect more effort has to be made on this subject. In our case the result indicates that neutral red might have a positive effect on the growth of our constructed E. coli strain.

Cyclic voltammetry - mediator characterization

The next improvement step in our experiments was to carry out cyclic voltammetry measurments. During our reasearch we found out that the workgroup from Dr. Dirk Holtmann of the Dechema research institute in Frankfurt investigates electroactive microorganisms. We visited them in their laboratory and gained lots of helpful recommendations. THey told us to use a potentiostat for further cultivations and helped us to organize one for our project.
A potentiostat balances the current flow and secures to work at a constant potential. It also functions as a measuring system and therefore provides a way to give a statement if the cells consume electric current.
For our experiments we used a Ag/AgCl reference electrode for measuring the working electrode potential. The counter electrode, which completes the cell circuit, was made from platinum wire. Platinum has the advantage that it is an inert conducter. The set up for the characterization of different mediators is shown is figure 6.

Figure 6: The H-cell reactor set-up for the characterization of different mediatory by cyclic voltammetry consists of the following parts: 1 Ag/AgCl reference electrode 2 Platinum working electrode 3 Platinum wire counter electrode 4 Heating water system.

The following figures show cyclovoltammogramms for neutral red (NR) and bromphenol blue (BPB) which are both promising candidates to function as mediators. The characterization was performed cell free in phosphate buffer.
The aim was to detect the potentials were the mediator gets reduced and where it gets oxidized again. We were interested in the potential where the mediator gets reduced, because that potential complies with the conditions where we need to carry out our cultivations.
Furthermore cyclovoltammogramms allow it to give a statement if the redox-reaction is reversible or not. For our measurements we varried the electrode material, the scan-rate, the step-size, the scan-limit and the positioning of the electrodes. In addition we investigated the influence of oxygen in the experimental set-up.
For oxygen free measurements we fumigated the cathode space with nitrogen for approximately 8 hours.

Figure 7: Cyclic voltammogramm with 100 µM bromphenol blue in phosphate buffer.

Figure 8: Cyclic voltammogramm with 100 µM neutral red in phosphate buffer.

Parameter	Value
Mediator	Neutral red
Scan rate [mV-s]	20
Step size [mV]	1
Scan limit E1 [V]	0.51
Scan limit E2 [V]	-0.8
Electrode material	Platinum
Aeriation	Oxygen is present

In figure 8 are two oxidations peaks present due to presence of oxygen. The small scan rate and step size result in a very flat curve. THe reduction peak is at a similar potential as it is in figure 7 which leads to the assumption that buffer components are reduced and oxidized. The potential where these reactions take place might overlax the reduction peak of neutral read. The redox potential of neutral red accounts to -325 mV vs. NHE.(Futz, M. L. & Durst, R. A., 1982)

Figure 9: Cyclic voltammogramm with 100 µM neutral red in phosphate buffer.

Parameter	Value
Mediator	Neutral red
Scan rate [mV-s]	10
Step size [mV]	1
Scan limit E1 [V]	0.1
Scan limit E2 [V]	-0.6
Electrode material	Platinum
Aeriation	Oxygen free by aeriation with nitrogen

Multiple measure cycles were performed to proof if the mediator properties change over the time. The measurement shows that the redox-reaction of neutral red seems to be stable.

Figure 10: Cyclic voltammogramm with 100 µM neutral red in M9 minimal media with 50 mM Xylose as carbon source.

Parameter	Value
Mediator	Neutral red in M9 minimal media- xylose (50 mM) with neutral red
Scan rate [mV-s]	10
Step size [mV]	1
Scan limit E1 [V]	0.7
Scan limit E2 [V]	-0.6
Electrode material	Platinum
Aeriation	Oxygen present

The measurement shown in figure 10 was performed in M9 minimal-media. We wanted to be sure that neutral red shows a similar redox-reactions in the media we used for the following cultivations.

Chronoamperometry - current consumption

All cultivations were carried out in the H-cell reactor at 37 °C and an air flow of 0.75 standard litres per minute. The used media was M9 minimal media with 50 mM Xylose as carbon source.
The technology of chronoamperometric measurements allows it to tell how much current was needed to keep a set potential constant. The amount of current that is necessary to hold the potential allows conclusion on the energ consumption of the cells.
An additionally performed analytical method was the GloAssay to detect the NAD⁺ and NADH levels in the cells. This should give hints to the state of the cell metabolism.

Figure 11: Cultivation of the E. coli KRX ΔdcuB::oprF strain in M9 minimal media- xylose (50 mM) with 100 µM neutral red added. During the cultivation there was set a potential of -400 mV on the H-cell achieved by the chronoamperometric method. The figure shows the optical density, Xylose concentration, and the NAD/NADH level during the cultivation, plotted against time.

Figure 12: Cultivation of the E. coli KRX WT in M9 minimal media with 100 µM neutral red added. During the cultivation there was set a potential of -400 mV on the H-cell achieved by the chronoamperometric method. The figure shows the optical density, Xylose concentration, and the NAD/NADH level during the cultivation, plotted against time.

Figures 11 and 12 should illustrate if neutral red influences the cell growth if an electrical current is set onto the cells. The results indicate that their is no difference in their growth characteristics. The levels of NAD and NADH seem to be constant in both cultivations.
To compare these results we cultivated the E- coli KRX wild type as a reference. The results are shown in figure 13.

Figure 13: Cultivation of the E. coli KRX WT in M9 minimal media with 100 µM neutral red added. During the cultivation there was set a potential of -400 mV on the H-cell achieved by the chronoamperometric method. The figure shows the optical density, Xylose concentration, and the NAD/NADH level during the cultivation, plotted against time.

The E- coli KRX wild type does not show significant changes in its growth characteristic in comparison to the cultivations in figure 11 and 12.
To proof if neutral red does really not influence the cell metabolism further experiments should be carried out. Additionally we were not able to generate biological replicates due to the fact that we were limited in H-cell reactor capacities.

Figure 14: Overlay of the chronoamperometric measurements from cultivation A B and C.

During the shown cultivations in figure 11 to 12 we performed chronoamperometric measurements. The results are shown in figure 14. All three curves are nearly identic which leads to the conclusion that there is no difference in their electron uptake. Further measurements might lead to a further understanding.

Flow Cell

The figures 15 to 17 show the set up and parts of our FCR. Unfortunately we were not able to use it, because we could't finish our cytochrome construct.

Figure 15: Basical set up of the FCR: 1 Hose pump 2 FCR connectet to an energy supply source and to the pump 3 Stock bottles for media and buffer.

Figure 16: Basical set up of the FCR: 1 Connection for the power supply cabel at the anode 2 Connection for the power supply cabel at the cathode 3 Hose connection nipple.

Figure 17: Individual components of the FCR: 1 Stable base for the cell 2 End cover plate with electrical plug for power supply 3 Separation partitions for anode- and cathode-space 4 Sealing rings 5 Screws for the fixation of the single parts 6 Carbonic electrode material.