Team:TU Delft-Leiden/Project/Life science/EET/characterisation

Module Electron Transport – Characterization

In the wet lab we integrated the Electron Transport pathway of Shewanella oneidensis into Escherichia coli. Here you can find information with respect to characterization of the BioBricks for the Electron Transport pathway.

Our BioBrick BBa_K1316012 encodes the mtrCAB genes under control of an weakened T7 promoter with the lac operator (T7 lacO). The MtrC, MtrA and MtrB proteins form a conduit to transfer electrons across the membrane. To complete the implementation of the extracellular electron pathway we constructed BBa_K1316011 as well , a BioBrick encoding the ccm genes under control of the pFAB640 promoter. The Ccm proteins help to mature the MtrCAB conduit. The elaborated combination of promoters and coding sequences that we used was found to generate the largest maximal current [1]. To test if this is true, we characterized our BioBricks. We made use of UV-vis and SDS-page to see if the proteins are expressed. In addition, we made use of a bioreactor and potentiostat to do advanced current measurements. In this way we were able to visualize expression of the ccm genes and showed that induction of our BioBricks indeed results in a current flow.

Characterization of BioBricks via Protein Determination

Our BioBrick BBa_K1316011 consists of genes encoding several cytochrome C maturation proteins (Ccm). The Ccm system mainly consists of heme delivery proteins that help the conduit proteins MtrCAB to mature properly. The Ccm proteins translocate heme in the periplasm and catalyze the formation of thioether bonds that link heme to two cysteine residues. The axial ligands are then coordinated to the heme iron and the holocytochrome C is folded. In E. coli C43 strains transformed with our Ccm BioBrick only, the Ccm proteins should be expressed in higher amounts compared to C43 strains who do not have the BioBrick. Increased expression of the Ccm proteins can easily be spotted the by redness of membrane pellets caused by the heme.

Membrane purification and UV-VIS

To look into the expression of the cytochrome c maturation (Ccm) proteins, an UV-VIS spectrum has been recorded. E. coli (C43) cultures were transformed with our BioBrick BBa_K1316011 and cultivated aerobically. E. coli (C43) strains transformed with the ccm & mtrCAB gene constructs described by Goldbeck et al. [1] and an empty E. coli (C43) strain were used as a positive and negative control respectively. Strains carrying BBa_K1316011 are expected to have more expression of the Ccm proteins than empty strains. More or less equal expression levels of the Ccm proteins are expected when strains carrying BBa_K1316011 are compared with E. coli (C43) maintaining both the ccm and mtrCAB clusters.

Before membrane protein purification, BBa_K1316011 and the controls where induced with IPTG to activate the promoters. Visual analysis of the pellets showed that there was a difference in color in strains stransformed with BBa_K1316011 and both the ccm & mtrCAB genes compared to empty E. coli (C43) bacteria. The BBa_K1316011 and E. coli (C43) Ccm+MtrCAB had a red color, which could be indicating the increased production of cytochrome c proteins because of the heme delivery proteins. Membrane protein purifications where done by low-speed and high-speed centrifugation.

According to [2], there should be a peak around 550nm for cytochrome c proteins. Using the UV-VIS results, there is a peak for all membrane fractions around 550nm, so it is possible to confirm the expression of cytochrome c proteins in all the samples. There is a difference between E. coli (C43) without plasmid and BBa_K1316011, as shown in figure 1. These observations confirm that BBa_K1316011 and E. coli (C43) Ccm+MtrCAB offer enhanced cytochrome c expression compared to the E. coli (C43) strain without the Ccm or MtrCAB plasmids.

SDS-PAGE

An SDS-PAGE has been done with the membrane fractions for E. coli (C43), E. coli (C43) Ccm+MtrCAB and BBa_K1316011. As mentioned before, no conduit proteins, such as the proteins in the MtrCAB operon, are included in BBa_K1316011, but they are included in E. coli (C43) Ccm+MtrCAB. When the MtrCAB operon is included, the membrane fractions of E. coli (C43) Ccm+MtrCAB should contain MtrA, a 32-kD periplasmic decaheme cytochrome c, MtrC is a 69-kD cell-surface-exposed lipid-anchored decaheme cytochrome c and MtrB is a 72-kD predicted twenty-eight strand β-barrel outer membrane protein.[1]

There are no significant differences between BBa_K1316011 and both positive and negative controls observed using SDS-PAGE. There are no MtrCAB proteins observed when looking into the SDS-PAGE lane for E. coli (C43) Ccm+MtrCAB. Due to the the weak MtrCAB promotor, the expression of MtrCAB proteins might be too low to be visible on the gel.

Measurements of EET in Self-Constructed Bioreactor

Shewanella oneidis MR-1 uses the MtrCAB proteins, the principal proteins in this module, to extracellularly reduce bulky metal oxide crystals which it uses as terminal electron acceptors in its respiration. Electrons stem from the intracellular oxidation of (organic) electron donors, and the process is thermodynamically favourable under physiological conditions. In this project we don’t seek to reduce metal-oxides but rather a working electrode in a three electrode cell.

Introduction to voltammetry

The three electrode cell is used to perform voltammetry which is an electro analytical method used to investigate the half-cell reactivity of an analyte. In voltammetry potential-difference (E) between a working and a reference electrode in an electrochemical cell is controlled and the resulting current (I) is measured. The working electrode is in physical contact with the analyte thereby facilitating the transfer of charge when a potential is applied. The reference electrode has a known, stable electrode potential and is used to gauge the potential of the working electrode. The third electrode is the auxiliary (or counter) electrode which balances the charge in the cell; it reduces or oxidizes any molecules that are in the solution. When no red-ox reactions take place at the working electrode, only a marginal current flows because of the applied potential between the reference and working electrode due to electrostatic effects. When the working electrode is either reduced or oxidized electrons flow through the circuit which can easily be detected using an Amperometer. In most voltammetric experiments the potential is varied at differing rates over time, however in this set-up the potential is kept constant for the course of the experiment. When a positive potential is applied to the working electrode in our set-up, electrons present on the extracellular side of the outer membrane of our engineered E.coli reduce it hence: a current flows. More on the potentiostat that our team can be found in the potentiostat subsection .

Our bioreactor

Figure 4 shows a schematic representation of our bioreactor. The working electrode is made of a square piece of carbon cloth (0.031m2)[REF] which is folded and tied together with a tie wrap to make it fit in the bioreactor. Carbon cloth has a large surface to volume ration, is non-toxic and therefore ideal for voltammetry when handling live organisms. The counter electrode is made of a graphite rod that is wrapped in silicon tubing to prevent any shorts due to the two electrodes touching. The reference electrode is silver/silver chloride (Ag/AgCl) with a saturated KCl electrolyte solution, yielding an electrode potential of Eref= +0.197 V versus a Standar Hydrogen Electrode (SHE)[3]. When a working electrode potential of for instance E = 0.2V is applied this means that the potential of the working electrode is actually E = 0.2 + 0.197 = 0.397V vs SHE. The temperature in the bioreactor is controlled through a heat mantle around the compartment where the cells are situated which is fed with warm water from a warm water-bath. The broth in the bioreactor is stirred with a magnetic stirrer, and there is a sampling tube present to take samples for OD600 measurements. Due to the nature of the cascade of reactions yielding the electrons that finally reduce the working electrode the broth needs to be completely anoxic, as pointed out by the modelling of the carbon metabolism [LINK]. To keep the broth free of oxygen a gas inlet is attached to a needle which feeds sterile N2 into the reactor close to the stirrer. To depressurize the reactor also a gas outlet is present. A picture of our bioreactor to which all above-mentioned components attached, and pictures of the individual components is shown in figure 5.

Metabolism and the source of electrons for the MtrCAB pathway

There is but a limited scope of substrates that can act as electron donors for the MtrCAB pathway which are: lactate, N- Acetylglucosamine, formate, and hydrogen [4]. In our experiments lactate is used as an electron donor since, when present at relatively high concentrations, it is dehydrogenated by lactate-dehydrogenase (LDH) to pyruvate and yielding NADH as seen in figure 6:

The NADH is then oxidized to produce menaquinol which then yields its electrons to the MtrCAB proteins via E. coli’s native NapC. Other carbon substrates like glucose ferment for which reason these substrates do not yield an excess of NADH which is essential for fuelling the MtrCAB pathway. To prove this principle we also used glycerol as a carbon source instead of lactate which can be fermented anaerobically, therefore theoretically yielding no electrons for the MtrCAB pathway. For more information on the carbon metabolism, see Carbon Metabolism and Electron Transport.

The experiments

In the first experiment we tried to roughly replicate the conditions as stated in the Jensen [1] article; the exact protocol for seeding the bioreactor can be found in the protocol for bioreactor here . Figure 7 represents the current in mA divided by the first OD600 measurement at the start of the experiment; figure 8 represents the OD600 measurements over time, for which the raw data can be found here . OD600 is not directly correlated to current, so only the first OD600 measurements is used to normalize the data for comparison. Cells in all experiments are grown in M4 minimal medium supplemented with 40mM D/L-lactate, except for one measurement where the cells were grown in M4 with 40mM of glycerol. E.coli C43 bearing the BBa_K1316012 (MtrCAB + T7 pLac) insert in a non-biobrick backbone and BBa_K1316011 (ccm cluster + pFAB640) in a non-biobrick backbone is referret to as the Ajo-F strain, and 'Empty' cells are non-transformed E.coli C43 which serve as a negative control. The bio-bricks were assayed in non-bio-brick backbones because these have a low copy-number, and cloning into biobrick vectors took a long time. Choosing a low copy number vector is crucial to the successful implementation of the pathway since the MtrCAB proteins become toxic when over-expressed. We implemented BBa_K1316012 (MtrCAB + T7 pLac) in a pSB3K3 vector which has a comparably low copy-number to the vector used in [1].

Figure 7A shows a significant difference in current measured between the empty C43 strain and the Ajo-F strain that was induced overnight at working electrode potential E = 0.2V. The difference is roughly 0.7mA at the beginning of the experiment, but this difference decreases over time. This decrease might be due to the faster drop in OD600 of the Ajo-F strain compared to the C43 strain. If faster decrease of OD600 were to be the explanation for decreasing current, it proves that the 'concentration' of cells is correlated to the observed current. Since the only difference in experimental procedure is the presence of the Ccm and MtrCAB proteins in the Ajo-F strain this result suggests that the observed current is indeed due to the functioning of these proteins. When the Ajo-F strain was re-suspended in M4 medium supplemented with 40mM glycerol it showed the exact same current as the empty C43 strain, proving that current was observed because of above-mentioned lactate dehydrogenation and subsequent steps leading to the excretion of electrons. 'Empty' C43 cells also yield a current, and this is due to various chemicals that cells produce which may be involved in red-ox reactions with the working electrodes.

Figure 7B shows that there is an even more significant difference in observed current between 'empty' C43 cells and the Ajo-F strain at working electrode potential E = 0.2V. This is of interrest because the aim is to make a biosensor that can produce quantative data, and therefore a larger contrast is more preferable. Once again a steep drop in OD600 values can be detected in the Ajo-F strain that was induced overnight as shown in figure 8. This drop in OD600 is again roughly correlated to the decrease in current, proving that current is produced by live cells. The reason that cells die at all when seeded in the bioreactor is probably due to the fact that they cant produce enough ATP to sustain their maintenance needs in terms of ATP production. Lactate is not a substrate that yields energy when catabolized anaerobically by bacteria.

In experiments with working electrode potential of either E= 0.2V or E = 0.4V the OD600 of 'Empty' C43 cells drops at a more gentle slope than that of C43 with induced MtrCAB proteins; this indicates that cells are dying at a slower rate. The difference in the rate of decline in OD600 values might be due to either or a combination of three reasons; the first reason being that MtrCAB proteins make the membrane more susceptible to tears caused by electrostatic effects. The second reason is that the counter electrode potential is fluctuating to more extreme values of potentials, again rupturing the cells. The third reason is that because of the dehydrogenation of lactate, toxic quantities of NADH build up inside the cell, eventually killing it.

As one might have noticed is that only negative controls ('empty' C43 cells) and the Ajo-F strain were tested; neither the uninduced Ajo-F strain nor C43 cells with the biobricks in a bio-brick (pSB) backbone were assayed. This is because the cloning of the BBa_K1316012 took so long that it was only finished two weeks before the wiki freeze. This indeed sounds like enough time to do the bio-reactor experiments; this was however not possible because the bioreactor broke, yielding us empty-handed. The strain containing the two aforementioned bio-bricks in pSB backbone was assayed with our micro-fluidics potentiostat system (Dropsens), as described below.

Conclusion and future experiments

In conclusion our biobricks themselves worked as expected yielding a current when a positive potential is applied to a working electrode of a potentiostat, and the current increases at larger working electrode potentials. Due to time-limitations we did not assay the inducability of the pathway, which would be crucial in terms of making a biosensor.

Future work with a bio-reactor should include the C43 strain with the biobricks in biobrick (pSB) vector. Also experiments where the cells are seeded into the bioreactor before being induced, and induction happening after seeding would aid the proof of the use of these biobricks as a biosensor output. Furthermore different media supplemented with various electron donors such as formate would be interesting, to see what electron donor/medium combo yields the best results in terms of current and cell-death. Also media with a mix of a C-source and an electron donor would be very interesting since the cells would than potentially grow, while also still fuelling the MtrCAB pathway with electrons.

DropSens as Microfluidic Device for Measurements of Current

The DropSens Microfluidic device described in the microfluidics section, was used to characterise the BBa_k1316011 and BBa_k1316012 biobricks. These constructs had already been characterised using the bioreactor, so the goal of these experiments was to establish whether the system could be scaled down to microlitre volumes.

6 different analytes were tested in the device: the induced and uninduced Arjo-Franklin strains were used as positive controls, and empty E. Coli chassis, and water were used as a negative controls. Finally, Induced and uninduced strains containing the two biobricks in question were tested. Water was flushed through the device between each test. These tests were all performed at 0.2v and 0.4v electrode potential, as was used in the bioreactor experiments.

At 0.2v, we see an exponential decrease in current after application of potential. This can be explained the a capacitance effect between the electrodes. All strains stabalise to a current approximately equal to zero - and less than that of pure water. It can be concluded that at the small volumes involved with the microfluidics device (33ul) there simply is not enough cells to produce a measurable current. Reasons for the negative currents seen in the Arjo-Franklin strains remain unknown at this time.

Similarly, at 0.4v electrode potential, a capacitance effect is seen, followed by no significant current from any of the strains. The flat line of the uninduced Arjo-Franklin suggests an equipment malfunction and thus should be discounted.

Conclusions

Measurable outputs from microfluidic devices with our cells is not possible, so further development, either in the biology, or a redesign of the device should be chosen - perhaps with an alternative solution to microfluidics.

This being said, the use of our designed microfluidic device offered several advantages over standard bench techniques (such as the bioreactor), so it is a field worth pursuing for use in other applications to improve the ease and efficiency with which experiments can be carried out.

References

1. C.P. Goldbeck et al., Tuning promoter strengths for improved synthesis and function of electron conduits in E. coli, ACS Synth. Biol. 2 (3), pp 150–159 (2013)

2. Reedy, C.J. & Gibney, B.R. et al., Heme protein assemblies, Chem Rev 104 (2), pp 617–49 (2004)

3. http://en.wikipedia.org/wiki/Standard_hydrogen_electrode

4. Frederik Golitsch, Clemens Bücking, Johannes Gescher, Proof of principle for an engineered microbial biosensor based on Shewanella oneidensis outer membrane protein complexes. Biosensors and Bioelectronics, 47, pp 285–291 (2013)