Aim

Establish a simple method to demonstrate that E. coli produces B12. Since there were no recipe for the ethanolamine medium that was used by Fowler et al. in 2010 [3] as well as the medium used by Chang & Chang et al. in 1975, available [1], the aim of our first experiment was to test the medium preparation and whether or not E. coli grows with ethanolamine as the sole nitrogen source without AdoCbl supplementation.

Experimental setup

E. coli BL21(DE3) were grown in a 5 mL overnight culture in LB and washed with DPBS. For this the cells were centrifuged for 5 min at 8000 rpm, supernatant removed and resuspended in 1 mL DPBS. The suspension was split and centrifuged again for 5 min at 8000 rpm. The supernatant was removed and the pellets resuspended in 100 mL either LB or ethanolamine medium in 250 mL erlenmeyer flasks. The cultures were incubated at 37°C and 200 rpm shaking. Samples are taken every hour up until the 8th hour with additional measurements at +12 hours and +24 hours and the absorbance at OD600 was measured using a ThermoScientific NanoDrop 2000c Spectrophotometer.

Results

While E. coli is able to grow in LB medium, shown by an increase in OD600, there was no growth observed in ethanolamine medium (Figure 8).

Conclusion

It is apparent that without adding AdoCbl to the ethanolamine medium E. coli is not able to grow. The next step is to show that the growth depends on the presence of AdoCbl.

Aim

To further prove that ethanolamine is the growth inhibiting factor in the medium we compare normal M9 medium with modified M9 medium where the Ammonium chloride is replaced with ethanolamine.

Experimental setup

We let E. coli BL21(DE3) grow in an overnight culture (5 mL) in LB and wash them twice with DPBS to remove any LB residue on the bacteria. For this the cells were centrifuged for 5 min at 8000 rpm and resuspended in 100 µL DPBS After the last centrifugation, the pellet was split and resuspended in 100 mL either M9 or modified M9, where the Ammonium chloride was replaced with ethanolamine, in 250 mL erlenmeyer flasks. The cultures were incubated at 37°C and 200 rpm shaking. Samples are taken every 12 hours for 48 hours and the absorbance at OD600 was measured using ThermoScientific NanoDrop 2000c Spectrophotometer.

Results

E. coli is able to grow in M9 medium but seems to stagnate at OD600 of around 1.5 (Figure 10). This is much lower than in LB medium (Figure 8), where OD600 values of 5 could be achieved. For the modified M9 medium with ethanolamine we saw no significant growth over 48 hours, as observed with ethanolamine medium (Figure 8) .

Conclusion

It is apparent that E. coli is not able to grow on ethanolamine as its sole nitrogen source. Therefore the next step will be to test if the cells are able to grow with ethanolamine medium supplemented with AdoCbl.

Aim

After establishing two versions of an ethanolamine medium, the next step is to show that E. coli can grow on ethanolamine as its sole nitrogen source when supplemented with AdoCbl.

Experimental setup

We let E. coli BL21(DE3) grow in an overnight culture (5 mL) in LB and wash them twice with DPBS to get rid of any LB residue on the bacteria. For this the cells were centrifuged for 5 min at 8000 rpm and resuspended in 100 µL DPBS After the last centrifugation, the pellet was split and resuspended in 5 mL either LB medium, ethanolamine medium, M9 medium or modified M9 medium where the ammonium chloride was replaced with ethanolamine.
AdoCbl was added to the media to a final concentration of 0 nM, 50 nM, 100 nM, 500 nM and 1 µM AdoCbl respectively. Reaction tubes were wrapped in aluminum foil to prevent AdoCbl degradation due to light exposure. The cultures were incubated at 37°C and 200 rpm shaking. Samples are taken every 12 hours for 48 hours and the absorbance at 600 nm was measured using ThermoScientific NanoDrop 2000c Spectrophotometer.

Results

E. coli BL21(DE3) grew in LB medium (Figure 12A) and M9 medium (Figure 12C) without a clear trend visible for the different AdoCbl concentrations.
Other than the OD600 in the LB medium making a uniform dip at 36 hours (Figure 12A).
For the modified M9 medium with ethanolamine we saw growth stagnation at an OD600 of around 0.1, whereas the setups with added AdoCbl reached an OD600 of around 0.6, with no apparent difference between the other setups (Figure 12D).
For the ethanolamine medium we observed a more pronounced trend with the added AdoCbl concentration. No added AdoCbl resulted in growth stagnation, whereas with increasing AdoCbl concentrations also the OD600 increased. Outlier being the setup with an AdoCbl concentration of 1 µM (Figure 12B).
When comparing the maximum growth between the different media we determined that the maximum OD600 reached for LB medium is around three times higher then the maximum in M9 medium. Moreover, cells grown in the modified M9 with ethanolamine reached half of the OD values compared to cells in M9 medium with AdoCbl supplementation (Figure 12E-I).
At an AdoCbl concentration of 50 nM the maximum OD600 reached in the ethanolamine medium is similar to the OD600 values reached by modified M9 medium with ethanolamine (Figure 12F). Supplementation with higher concentration of AdoCbl of 100 nM, 500 nM and 1 µM resulted in a similar growth behavior as seen in the M9 medium (Figure 12G-I).
The bacteria, in both media containing the ethanolamine, show a lag phase of around 24 hours.

Conclusion

The data must be interpreted with caution since the experiment was only done once with no replicates and the culture volume of 5 mL in reaction tubes being very low.

However, the data suggests that the modified M9 medium with ethanolamine as its sole nitrogen source could only be used as a qualitative way of determining AdoCbl content in a given medium.
On the other hand, we have shown that AdoCbl is the limiting factor in the breakdown of ethanolamine into ammonia for use in amino acid synthesis.
Modified M9 medium with ethanolamine did not show any significant growth at 48 hours whereas ethanolamine medium does, this led us to go further with this medium as our detection medium. We therefore assume that we can use this medium as quantitative measurement for AdoCbl and will be tested next.
After interviews with experts and professors it was suggested to us by Prof. Boll that it could be very stressful for E. coli when growing in LB and then being transferred to a minimal media containing ethanolamine. If you are interested in learning more about our meeting with Prof. Boll or other meetings we had with academic personnel or industry stakeholders, take a look at our Integrated Human Practices page. Since E. coli would need to adapt its whole metabolism to the new medium. It was therefore suggested to us we let E. coli adapt in ethanolamine medium first with very low amounts of AdoCbl added.
Therefore, it is not surprising that 48 hours is not enough to observe growth stagnation. As a result, the next experiments will be conducted for a longer period of time (72 to 96 hours) to measure the maximum OD600 possible in the ethanolamine medium.

Aim

To test whether maximum OD600 values in ethanolamine medium allow conclusions to be drawn about the AdoCbl content in the medium, we aim to establish a calibration curve for AdoCbl content added to the medium.

Experimental setup

We let E. coli MG1655 containing the pGGAselect plasmid grow in a LB overnight culture (100 ml, at 37°C and 200 rpm) with added chloramphenicol (34 µg/mL).
After 12 hours the bacteria were centrifuged for 6 minutes at 4500 rpm and washed twice with DPBS and resuspended in ethanolamine medium. Then they were transferred to 100 mL ethanolamine with 1 nM added AdoCbl and grown for 24 hours at 37°C and 200 rpm. After 24 hours the culture is centrifuged for 6 minutes at 4500 rpm washed once with dPBS and then distributed to the different setups to a starting OD of around 0.1.
Samples are taken every 12 hours for 96 hours and OD600 is measured using the SpectraMax ID5 plate reader.

Results

For the setup with 0 nM and 1 nM AdoCbl we saw no significant growth after 72 hours (Figure 14). Whereas for the setups with 50 nM, 100 nM, 500 nM and 1 µM AdoCbl added to the medium we saw a gradual increase of maximum OD600 at 72 hours. But all of them are still within the previous error ranges (Figure 14). Interestingly for the setup with 2 µM AdoCbl added we see a decrease in maximum OD600 compared to the setup with 1 µM AdoCbl.

Conclusion

It is apparent that a minimum concentration of AdoCbl added to the medium is necessary for E. coli to grow in the ethanolamine medium which seems to be somewhere between 1 nM and 50 nM.
A smaller effect can be seen with increasing AdoCbl concentration in the medium, but the main factor contributing to the growth would be the AdoCbl uptake. This is due to the expression of the corrin ring transporter BtuB being downregulated by a riboswitch upon binding to AdoCbl, resulting in lower uptake when AdoCbl is already in the cell [4], [5].
This allows no conclusions to be drawn to the actual AdoCbl concentrations inside the cell. And are most likely very similar between the different setups even though we would expect that since AdoCbl is used in EAL that it wouldn't be available to the riboswitch and therefore allowing more AdoCbl to be taken up.

Aim

It has been shown that E. coli is able to produce low levels of AdoCbl when supplemented with Cbi since low amounts of DMB are available in the cell [3].
We want to test this by supplementing E. coli with different Cbi concentrations. This will furthermore help us to determine if the added amount of Cbi (500 nM) is enough for E. coli to produce any meaningful amounts of AdoCbl without introducing BluB into the organism.

Experimental setup

E. coli BL 21 (DE3) grow in an overnight culture (5 mL) in LB and washed them twice with DPBS to get rid of any LB residue on the bacteria. For this, the cells were centrifuged for 5 min at 8000 rpm and resuspended in 100 µL DPBS. After the last centrifugation, the pellet was split and resuspended in 4 mL ethanolamine medium in reaction tubes.
The substrate Cbi was added to the media to a final concentration of 0 nM, 500 nM, 1 µM, 5 µM and 10 µM Cbi respectively. The cultures were incubated at 37°C and 200 rpm shaking.
Samples are taken every 12 hours for 96 hours and the absorbance at OD600 was measured using SpectraMax ID5 plate reader.

Results

When comparing E. coli's growth it shows no significant increase in OD600 for 0 nM, 500 nM and 1 µM Cbi over 96 hours (Figure 16). Supplementation with 5 µM and 10 µM Cbi lead to an increased growth, however, with OD600 values lower than observed for cultures supplemented with 50 nM AdoCbl (Figure 14).

Conclusion

We conclude that adding the production amount of 500 nM Cbi has no influence on the ethanolamine bioassay, as basal AdoCbl production appears to be insufficient to support growth in ethanolamine medium. Only with higher Cbi concentration (5 and 10 µM) cells are able to produce enough AdoCbl to grow. This could be due to higher diffusion leading to more uptake of Cbi into the cell overcoming native transporters [6]. Note that these concentrations do not necessarily correspond to intracellular concentrations.

Key Findings

We developed a biosensor that indicates the presence of AdoCbl by fluorescence.
Since the riboswitch inhibits transcription/translation of downstream genes once AdoCbl binds to it, we should detect a decrease in a fluorescent signal when we would simply combine the riboswitch with a fluorescent protein.
For us, this did not seem like the best solution for a readout. Inspired by the work of Yingying Cai et. al. [7] and iGEM Wageningen 2016, our solution is to control the expression of the LacI repressor by the riboswitch (Figure 17). With this, we can produce a positive readout for the activity of the riboswitch.
Presence of AdoCbl would result in an increased expression of a fluorescent protein (e.g. superfolder GFP).
We choose three different riboswitches to test. One originated from E. coli K12, one from P. freudenreichii and the last one from S. typhimurium, referred to as K12 and PF and SY, respectively.

Aim

Explore in what range of AdoCbl concentrations the riboswitch works specifically.

Experimental setup

E. coli MG1655 containing the pIG_K12BS and pIG_PFBS plasmids, are taken from a glycerol stock and added to 5 mL M9 medium with spectinomycin (50 µg/mL) for the sensors and Chloramphenicol (34 µg/mL) for the pGGA.
Different AdoCbl concentrations are added (0 nM, 10 nM, 100 nM, 200 nM, 500 nM, 1 µM and 2 µM AdoCbl). As a positive control we induce one setup with 200 µM IPTG. As a negative control E. coli MG1655 containing an empty pGGAselect plasmid was tested. For better comparison we also included a sensor without a riboswitch (piG_04).
Fluorescence and OD600 were measured after 12 hours using the SpectraMax ID5 plate reader. Fluorescence was normalized to OD600.

Results

For the K12 sensor, we observed an increase of fluorescence signal for the AdoCbl concentrations: 1 nM, 10 nM and 100 nM (Figure 19A), while there was fluorescence stagnation at higher AdoCbl concentrations.
We also observed an increase of fluorescence signal for the PF sensor, but there it stagnates already at 10 nM AdoCbl, with no noticeable change at higher concentrations. (Figure 19B).
It should be noted that a high background fluorescence was observed for both riboswitches, as seen with the 0 nM AdoCbl (Figure 19). This is higher than the controlled autofluorescence of cells expressing pGGAselect (Figure 19). The overall increase of around 1.8E+7 AU seems to be the same for both sensors when comparing 0 nM with 100 nM for the K12 sensor and 0 nM with 10 nM for the PF sensor (Figure 19A & 19B).
When 200 µM IPTG was added we saw a noticeable increase in fluorescence for both sensors with the K12 sensor being slightly higher than the K12 sensor and 0 nM with 10 nM for the PF sensor (Figure 19C). However, the fluorescent signal does not reach the values observed for the induced control (p_04, Figure 19C), which is around 2-3 times higher.

Conclusion

Since we observed an increase in fluorescence, albeit a rather small one, with rising AdoCbl concentrations, we conclude that these riboswitches and therefore our sensors function as intended, by repressing the repressor LacI leading to higher sfGFP expression.
However, we also noticed that both riboswitches do not appear to tightly regulate the repressor as even at saturating levels (>100 nM) of AdoCbl the fluorescence does not reach the fluorescence values as observed upon induction with IPTG.
This could be either due to the riboswitch still allowing lacI production. Or LacI being produced before saturating levels were reached in the cell and staying for longer time then 12 hours, still repressing the Trc promoter.
We assume that the riboswitch with its own ribosome binding site (RBS) does not provide adequate repressor function for LacI expression. This is further supported by results we gained through a toxin experiment, which also showcased a change in production after the RBS was replaced. More on this experiment is available on our Toxin-Antitoxin Results page.

to ensure tight repression of the reporter gene sfGFP, leading to a high background fluorescence. The next step would therefore be to either replace the RBS with a stronger one or add an additional strong RBS behind the riboswitch. It is not possible to derive parameters for the actual concentrations in which these riboswitches operate since we only know the AdoCbl concentration we provide in the medium. The actual AdoCbl concentrations in the cell are unknown. This will be tested in the next steps.

Aim

To have more data on different riboswitches which we could use to better design CELLECT we choose to test another riboswitch derived from S.typhimurium, referred to as SY riboswitch. During cloning and subsequent screening, it became apparent that after induction with 2 µM AdoCbl on the plate the colonies did not fluoresce under UV light even though sequencing confirmed correct assembly of the sensor. We decided to test it regardless and aim to explore the working range of the riboswitch by testing different AdoCbl concentrations.

Experimental setup

E. coli MG1655 containing the sensor is taken from a glycerol stock and added to 5 mL M9 medium with spectinomycin (50 µg/mL) in reaction tubes. Different AdoCbl concentrations are added (0 nM, 10 nM, 200 nM and 1 µM AdoCbl). As a positive control we induce one setup with 200 µM IPTG. As a negative control E. coli MG1655 containing the p_04 template plasmid was tested.
Fluorescence and OD600 are measured after 12 hours using the SpectraMax ID5 plate reader. Fluorescence is normalized to OD600.

Results

When comparing the different fluorescent values there is no clear trend visible and all values fall roughly in the same range.
Furthermore, we saw that the fluorescence for the SY sensor is very low even compared to autofluorescence of pGGA (Figure 21A). The SY sensor when induced with 200 µM IPTG, however, showed a fluorescence similar to the other sensors with the same induction (data not shown).
Moreover, we noticed that the K12 sensor and the K12 sensor and 0 nM with 10 nM for the PF sensor exhibit 20 times higher background fluorescence compared to the SY sensor (Figure 21B).

Conclusion

With the data, we conclude that the sensor does not work properly as expected. Possible reasons are not known at the moment and would need further testing. We decided to not continue working with that sensor.

Aim

After encountering high background fluorescence with the first versions of the sensors, as well as a low maximum fluorescence in the presence of AdoCbl we will test a new version of the K12 sensor.
Our hypothesis is that the native ribosome binding site (RBS) in the K12 riboswitch is too weak for a tight repression of the reporter gene. Therefore a new sensor was constructed with a stronger RBS (piG_K12BSb). This new sensor will be tested and categorized with a calibration curve over different AdoCbl concentrations.

Experimental setup

E. coli MG1655 containing the sensors are taken from a glycerol stock and added to 5 mL M9 medium with spectinomycin (50 µg/mL) for the sensors and Chloramphenicol (34 µg/mL) for the pGGA, in reaction tubes. Different AdoCbl concentrations are added (0 nM, 10 nM, 100 nM, 200 nM, 500 nM, 1 µM and 2 µM AdoCbl).
As a positive control we induce one setup with 200 µM IPTG. As a negative control E. coli MG1655 containing an empty pGGAselect plasmid was tested. For further comparability we also include two setups with the sensor that does not contain a riboswitch (piG_04) and induce one setup with 200 µM IPTG as well.
Fluorescence and OD600 are measured after 12 hours using the SpectraMax ID5 plate reader. Fluorescence is normalized to OD600.

Results

We observe a rather small but still steady increase of fluorescence with increasing AdoCbl concentration until a concentration of 500 nM, followed by a stagnation in fluorescence (Figure 23). Interestingly, this time we observed a higher background fluorescence with 0 nM AdoCbl (Figure 23A). However, K12b showed reduced fluorescent values compared to the tested K12 sensor.
Comparison between the K12 sensor and the K12 sensor with the new RBS (K12b sensor) reveals a twice as high background fluorescence as well as a twice as high fluorescence when IPTG is added (Figure 23B).

Conclusion

The steady increase of fluorescence can be attributed to the added AdoCbl.
But still the same problem persists as with the K12 sensor and the K12 sensor and 0 nM with 10 nM for the PF sensor: the increase in fluorescence observed is very low and insignificant when compared to fluorescence values obtained by supplementation with IPTG (Figure 23B).
It is worth noting that the K12b sensor showed much less background fluorescence than the K12 sensor. This suggests that more repressor is produced leading to a tighter regulation of the reporter gene.
An identical effect we were able to observe in a toxin experiment, you can find more on Toxin-Antitoxin Results page.
Therefore we conclude that placing a stronger RBS behind the riboswitch leads to higher protein expression without compromising the regulatory function of the riboswitch. This is due to the transcription regulational part of the riboswitch, since the sequestering of the RBS does not apply for the changed RBS with its different sequence.

Aim

Next we want to test if the sensors is able to detect Cbi, as well as to determine basal AdoCbl production. This will be important for the use of the sensors as a detection method, since the substrate we add should not be detected by the sensors. This will be done with the K12 and K12 sensor and 0 nM with 10 nM for the PF sensor as well as the K12b sensor with the new RBS.

Experimental setup

E. coli MG1655 containing the sensors are taken from a glycerol stock and added to 5 mL M9 medium with spectinomycin (50 µg/mL) for the sensors and Chloramphenicol (34 µg/mL) for the pGGA, in reaction tubes. Different Cbi concentrations are added (0 nM, 10 nM, 100 nM, 200 nM, 500 nM, 1 µM and 2 µM AdoCbl). As a negative control E. coli MG1655 containing an empty pGGAselect plasmid was tested.
Fluorescence and OD600 are measured after 12 hours using the SpectraMax ID5 plate reader. Fluorescence is normalized to OD600 to ensure comparability.

Results

When comparing the different Cbi concentrations tests, we see no clear trend and every value for the three sensors seems to fall into the error range (Figure 25A,B & C). But all values are higher than the autofluorescence of E. coli containing the pGGA plasmid (Figure 25).

Conclusion

From this we conclude that Cbi can not be detected by the sensors in the range we induce our production cultures (500 nM cbi). Furthermore, we conclude that no significant AdoCbl production is happening due to low levels of DMB in the cell. This is in agreement with the findings of Basal AdoCbl production from the ethanolamine assay.

Aim

Establish AdoCbl uptake of E. coli in M9 medium, as it would allow us to connect fluorescence reading with actual intracellular AdoCbl concentrations. And therefore give us more accurate ranges for the functionality of our sensors as quantitative ways of determining AdoCbl in a given bacteria.

Experimental setup

E. coli MG1655 with the pGGA select plasmid were grown in M9 medium overnight and then transferred to reaction tubes with 5 mL M9 added with chloramphenicol (34 µg/mL) to a starting OD600 of around 0.2.
The different setups for M9 medium are as follows: 0 nM AdoCbl, 500 nM AdoCbl, 1 µM AdoCbl. Two biological replicates will be done At each time point (+12 hours & +24 hours) we will take samples for LC-MS and OD600 measurements. Before measuring the samples all AdoCbl gets converted to OHCbl under brigth light for 2 hours. Since OHCbl is a more stable form of AdoCbl.

Results

No siginificant difference in OD600 is observed for all three setups over 24 hours (Figure 26).
We observe an increase of intracellular OHCbl concentration for both setups as well as both timepoints after 0, with added AdoCbl in the medium (Figure 27).
The OHCbl concentraitions raises for the setup with 1 µM OHCbl as compared to the setup with 500 nM OHCbl, where it decreases at 24 hours (Figure 27).
It is interesting to note that the OHCbl concentration for the setup with 500 nM AdoCbl added at 12 hours is the highest value observed and then dropping to half that at 24 hours (Figure 27).

Conclusion

From the high concentration at 12 hours as seen in the setup with 500 nM AdoCbl we conclude that that's about the limit of AdoCbl uptake, with no further uptake after that in that setup.
As the OD600 doubled the intracellular concentration halved suggesting that there was no further uptake, just dilution due to cell division. For the setup in 1 µM AdoCbl it seems that the diffusional pressure is great enough [6] allowing AdoCbl to be taken up even when the active transport is no longer used by E. coli.