Part:BBa_K4604023
piG_21 (tetR_bluB_riboK12_mazF_AmpProm_mazE)
BioBrick piG_21 is a unit of expression consisting of the tet promoter/repressor, bluB, an AdoCbl riboswitch, MazF, the rrnB terminator, the Amp promoter, MazF and the rrnB terminator. The backbone we used in the experiments is pGGAselect.
Usage and Biology
So far, antibiotic resistances have played a crucial role in genetic engineering, ensuring plasmid retention. Intriguingly, plasmid retention does not guarantee plasmid encoded gene expression, since bacteria can use methods such as methylation or deletion to withdraw from the metabolic burden. This means that bacteria have their ways of not expressing inserted genes, defying the researchers control. This composite biobrick represents a novel tool to ensure not just the presence of a plasmid but also the expression of recombinant genes on it in bacteria. The key idea is to couple the production of a metabolite to the survival of the cell and thereby making it autoregulatory.
Modularity
Our composite part incorporates a riboswitch (BBa_K4604031), a toxin-antitoxin system (BBa_K4604037/BBa_K4604011) and an enzyme (BBa_K4604005) needed to boost the Adenosylcobalamin (AdoCbl) production in E. coli. Together, these parts were designed to form an autoregulatory circuit regulating the survival of the engineered cells. The system was tested and developed with AdoCbl production as a proof of principle, letting only AdoCbl producing cells survive. However, this biobrick can be easily adjusted for other purposes where plasmid stability plays an important role. The adjustability originates in the possibility to change the riboswitch (for sensing the desired compound) and the bluB gene (essential for bioproduction). The target compound can be any small molecule either arising from degradation or production, as long as a riboswitch for it exists, which allows the system to detect its presence. The limitation for the need of an already existing riboswitch could possibly be overcome by the use of a synthetic riboswitch made on the basis of an aptamer [1]. Synthetic aptamers can be generated through in vitro systematic evolution of ligands by exponential enrichment (SELEX) which allows for an even broader field of application of the composite part. However, this is a complex process and requires extensive testing which was outside of the capabilities of our iGEM project. The bluB gene can similarly be exchanged with an enzyme that fits the chosen product. The host organism is also variable; even though the biobrick is adapted to use in E. coli, with adjustments of the promoter-/terminator region and the toxin-antitoxin system it could be applicable to any chassis organism. To explore this opportunity we investigated the feasibility of producing AdoCbl in cyanobacteria. Read more on our results in cyanobacteria on the results page https://2023.igem.wiki/freiburg/production-results.
Figure 1: Scheme of the final system
AdoCbl production
We decided on the AdoCbl (a bioavailable form of vitamin B12) production as a proof of concept. Vitamin B12 is an essential nutrient, humans are dependent on for the production of red blood cells, the synthesis of the DNA and the function of nerves. To form the complete AdoCbl synthesis pathway in E. coli, it would require 28 additional genes. Since this is not realistic nor practical for an iGEM project, we decided on an alternative method. When supplemented with cobinamide, a precursor for AdoCbl, E. coli is capable of producing AdoCbl on their own in small amounts. With the overexpression of a naturally occurring gene of the synthesis pathway, called bluB, a greater yield can be achieved [2].
Proof of the production
To characterize the functionality of this part we first of all used Western Blot, an ethanolamine medium and mass spectrometry to qualitatively and quantitatively prove the production of AdoCbl in E.coli MG1655.
Western Blot to demonstrate production relative to inducer strength
The Western Blot analysis using an anti-His-Tag antibody confirmed the induced expression of bluB. Different concentrations of the inducer doxycycline were tested to identify the optimal yield of the BluB protein.
PICTURE WESTERN BLOT (noch hochladen) Figure 2: BluB enzyme production for different inducer concentrations.Detection of recombinantly expressed, his-tagged BluB enzyme with SDS-PAGE followed by Western Blot. Detection of the BluB protein was performed with an anti-his antibody. Loading control: RNA polymerase β-subunit. E. coli BL21 DE3, [piG_01a, BBa_…] overnight culture in LB medium, uninduced.
While we observed an increase in Blub protein yield with increasing inducer concentration, we also noted a leaky expression without induction. This is most likely due to a silent mutation that we introduced in TetR to conform to iGEM standards for this part.
Ethanolamine Medium an easy method to check for AdoCbl production
Ethanolamine medium is a minimal medium devoid of any nitrogen source despite ethanolamine. Since nitrogen is crucial for cell division, E. coli is unable to grow in such a medium. Use of ethanolamine as a carbon or nitrogen source requires AdoCbl, which is the cofactor of ethanolamine ammonia lyase that catalyzes the conversion of ethanolamine to acetaldehyde and ammonia [3]. Cells can grow on ethanolamine only if there is AdoCbl available. Based on a publication from 1976 [4] we were able to produce a minimal medium in order to demonstrate the production of AdoCbl by overexpression of bluB. In this assay we compared the growth of piG_01b / BBa_K4604015 to the mutated non-functional bluB expression construct (piG_07 / BBa_K4604020) after induced production in M9 medium.
Cells were cultivated in M9 medium, induced with 100 ng/ml DOX and supplemented with the essential substrate for AdoCbl synthesis. After 24 hours the cultures were washed and then cultivated in the ethanolamine medium to observe growth.
PICTURE ETHANOLAMINE MEDIUM GROWTH (hochgeladen)
Figure 3: E. coli MG1655 growth curve comparison with maximum after 72 hours in 1975 ethanolamine medium. OD 600- measurement of culture samples containing plasmids pGGAselect, piG_01b or piG_07 using SpectraMax ID5 plate reader. The data present in these graphs is the result of at least two independent biological replicates.
The growth curves clearly show only cells capable of metabolizing the ethanolamine by producing AdoCbl are the induced ones that contain the functional bluB enzyme (piG_01b). E. coli are able to produce AdoCbl from DMB and cobinamide alone, but only in small amounts since they do not produce enough bluB to synthesize large amounts of DMB. With added DMB or an overexpression of BluB higher yields of AdoCbl can be obtained. This proves that bluB over-expression in E. coli leads to the AdoCbl synthesis crucial for ethanolamine metabolism.
Liquid Chromatography Mass Spectrometry (LC-MS) for quantitative testing
LC-MS is a method used to determine the concentration of molecules based on their mass. With this, it was possible to detect how much Hydroxocobalamin (OHCbl) we produce (relative to dry cell mass). OHCbl is another derivative of B12 that forms from AdoCbl when exposed to light. To make the preparation of the samples and measurement easier, we just worked in sunlight, let the AdoCbl turn into OHCbl and measured the amount of that. We cultivated bacteria containing either a functional or a mutated version of the blub gene and afterwards sent them to be measured with mass spectrometry.
PICTURE LC MS piG_01b/piG_07
Figure 4: OHCbl content measurement with LC-MS in dry cell pellet. E. coli MG1655 cultures induced with 100 ng/mL doxycycline were supplemented with 500nM cobinamide. Samples taken immediately before induction and after cultivation for 24 h. LC-MS performed at Hannibal Lab, University Medical Center Freiburg.
These results clearly show that the AdoCbl yield in our production culture is heavily influenced by the overexpression of a functional BluB and also quantifies how much AdoCbl we produce. It’s no wonder that compared to highly calibrated systems [5] we produce less AdoCbl (530.29 μg/g dry cell weight vs. ~47.4 µg/g dry cell weight), as this is heavily influenced by the circumstances. Of course, there is no denying that optimization of the conditions are necessary to achieve optimal results.
All these results state clearly that we were successfully able to achieve a production of AdoCbl in E. coli. See more detailed data under https://2023.igem.wiki/freiburg/production-results.
Riboswitch
Riboswitches occur naturally as regulators of gene expression. Riboswitches are encoded as DNA sequences which exert their regulatory effect upon transcription by folding into a complex 3D structure. Generally, a riboswitch can fold in such a way that the ribosome binding site (RBS) is available or unavailable for the ribosome. Upon binding to the target compound, like AdoCbl, a configurational change in the riboswitch changes the availability of the RBS. In E. coli a riboswitch regulates the expression of the btuB gene which encodes for a corrine transporter protein BtuB. It has been shown that when sufficient AdoCbl is present the riboswitch undergoes a conformational change resulting in lower btuB expression [6]. We decided on using this riboswitch to downregulate the toxin when AdoCbl is produced in sufficient amounts. To verify the performance of the riboswitch we created a sensor with a fluorescent readout inspired by the iGEM team Wageningen 2016 (BBa_K1913011). However, if we had placed the sfGFP (BBa_K4604034) directly after the riboswitch, its negative regulation in the presence of AdoCbl would have resulted in a decrease in fluorescence. In our biosensor (BBa…) the riboswitch is placed in front of a repressor gene (lacI) which inturn suppresses sfGFP expression. If the riboswitch is triggered by binding of AdoCbl, the repressor expression is stopped leading to a detectable fluorescent signal.
PICTURE SENSOR SCHEME
Figure 5: Biosensor for the detection of AdoCbl.By placing the negative controlling riboswitches in front of the repressor gene LacI a positive signal to AdoCbl inside the cell can be achieved
The biosensor was tested independently of AdoCbl production. For this reason, we added different concentrations of AdoCbl to the cultures and tested the sensitivity of the sensor by fluorescent measurement. Theoretically, AdoCbl is taken up by the cells, binds to the riboswitch and enables the expression of sfGFP.
GRAPH SENSOR FLUORESCENCE (konstantin lädt hoch)
Figure 6: Fluorescence measurement of E. coli MG1655 containing piG_K12BSb in M9 medium with different AdoCbl concentrations after 12 h incubation time.Fluorescence of cells containing pGGAselect was measured. Fluorescence measurement was done using SpectraMax Plate Reader.The data present in these graphs is the result of at least three independent biological replicates.
Fluorescent measurement, as shown in figure 6, indicates that at a concentration of 100 nM AdoCbl the riboswitch is saturated and can no longer recognize an increase of AdoCbl. This data verifies the sensibility of the riboswitch; it can precisely detect AdoCbl in different concentrations and accordingly regulate the gene downstream of it. However, we also noted a high background fluorescence without the addition of AdoCbl.
Toxin-Antitoxin System
Toxin-antitoxin systems (TA-systems) play a crucial role in plasmid stability for naturally occurring plasmids [7]. Usually, the toxin targets essential cellular functions and causes growth arrest or cell death, to which the antitoxin acts as a counterpart. Toxin and antitoxin exhibit differences in their stability and lifespan [8]. While the antitoxin has a shortened lifespan due to its sensitivity to degradation, the toxin has a longer lifespan and is more stable. If the plasmid that contains the TA-system is lost, the antitoxin is rapidly degraded and the toxin concentration increases, leading to cell death. Therefore, when first discovered, TA systems were called “addiction modules” that ensure plasmid retention.
MazE/F as an executer
The toxin-antitoxin system we chose is MazE/F. The labile MazE (antitoxin) acts as a neutralizer to the stable MazF (toxin), which is an endoribonuclease. When MazF is present freely in the cell it cuts cellular RNA which ultimately leads to cell death. It’s important to note that this system is already present and constantly expressed in the strain we use, therefore a baseline expression of the native system is given. The endogenous MazE/F system regulates its own gene expression [9]. We therefore decided to use a different promoter than the native one to prevent interference by the toxin-antitoxin complex. To gain insights on the functionality of the antitoxin-toxin system we performed several experiments.
Proof of inhibitory effect of MazF (expression) on cell growth
First, we needed to verify the functionality of the toxin MazF, for this purpose it was tested separately from the AdoCbl production. In a separate construct, namely piG_23a (BBa_K4604024), we identified the time span in which the toxin MazF kills the cells and confirmed its toxicity. This was accomplished with an assay consisting of OD measurements and colony-forming-units (CFU). Cell toxicity was tested in a liquid culture with different inducer concentrations. First a western blot was performed to detect MazF expression.
WESTERN BLOT MAZF
Figure 7: MazF toxin expression for different inducer concentrations. Detection of recombinantly expressed, FLAG-tagged mazF toxin with SDS-PAGE followed by Western Blot using anti-Flag antibody. Loading control: RNA polymerase subunit. Western Blot with samples from different time points of pGGAselect in MG1655 in M9 medium, induced and uninduced in Lämmli buffer; Western Blot with samples from different time points of Pl.23 in MG1655 in M9 medium, induced in non-degrading buffer.
Induced expression of mazF was confirmed by Western Blot. The bands at 15 kD are the monomeric mazF. Since mazF also forms an unstable dimer, the faint bands at a height of 30 kD probably are the mazF dimer. Moreover, we have observed an accumulation of mazF with increasing culture time. To determine the toxicity of the expressed mazF, we performed a growth assay of 48 hours (fig. 8).
OD GRAPH PL23A
Figure 8: E. coli growth curve comparison over 48 hours in LB medium, containing pGGAselect or piG_23a with different doxycycline concentrations. OD 600- measurement of culture samples using ThermoScientific NanoDrop 2000c Spectrophotometer.
CFU GRAPH PL23A
Figure 9: E. coli MG1655 colony forming units (CFU) assay. CFU/mL values for each timepoint for piG_23a and control pGGAselect.
A clear growth inhibition was observed in cultures of piG_23a induced with doxycycline concentration of 50 ng/ml and higher. To make sure that the toxicity is not due to the antibiotic nature of our inducer we tested different concentrations of doxycycline and their effect on cell growth.
OD GRAPH DOXY
Figure 10: E. coli MG1655 growth curve comparison in LB medium over 20 hours with different DOX concentrationsOD 600- measurement of culture samples using ThermoScientific NanoDrop 2000c Spectrophotometer.
After this we came to the conclusion that concentrations of up to 100 ng/mL are fine to use. These results clearly indicate the toxic effect MazF has on cells and their proliferation.
MazF regulated by the riboswitch
The next step was to test mazF in combination with the riboswitch to test if it can successfully regulate the toxin expression. For this, we conducted a toxicity assay via OD measurements and CFUs. As opposed to the experiment shown above this had to be done in M9 medium to prevent the riboswitch from being triggered by the AdoCbl concentration in LB-Medium. This caused a decreased growth rate.
OD GRAPH PL3
Figure 11: titelBildunterschrift, OD 600- measurement of culture samples using ThermoScientific NanoDrop 2000c Spectrophotometer.
CFU GRAPH PL3
Figure 12: titelBildunterschrift, CFU/mL values for each timepoint for piG_03 and control pGGAselect.
These graphs indicate inhibited toxicity of MazF if regulated by the riboswitch. Our first guess was that cells mutated the toxin due to the evolutionary pressure a toxin implies. However, after sending several samples of our cultures to sequencing, showing no mutations of the toxin, this thought was discarded. Consequently we kept searching for a cause. The inhibited expression hinted at the ribosome binding site of the riboswitch being too weak for adequately high toxin production. This was further supported by the fact that the biosensor is leaky even in high concentrations of AdoCbl. This indicated that the repressor was being expressed in concentrations too low to sufficiently inhibit GFP translation. Since we use different promoters in these plasmids we identified the RBS as the probable issue since it is the only regulatory element these plasmids share.
Decreased toxicity caused by RBS inherent in riboswitch
To investigate this hypothesis we modified the plasmid that proved to lead to cell death (piG_23a) by adding the ribosome binding site of the riboswitch in place of the present RBS (piG_23b). A toxicity assay was performed again.
OD GRAPH PL23B
Figure 13: titelBildunterschrift, OD 600- measurement of culture samples using ThermoScientific NanoDrop 2000c Spectrophotometer.
CFU GRAPH PL23B
Figure 14: titelBildunterschrift, CFU/mL values for each timepoint for piG_23B and control pGGAselect.
As shown in the graphs above the inhibitory effect of mazF is significantly decreased by adding the supposedly weaker RBS to the plasmids. This realization was a huge step forward in our research, yet we still had to find a way to make mazF work with the riboswitch. Unfortunately we ran into time limitations and could not successfully test the riboswitch with the stronger RBS and toxin in combination.
For more explicit information about toxin results see https://2023.igem.wiki/freiburg/toxin-antitoxin-results.
Conclusion/Future outlook
As shown by all the results the parts all by themselves work considerably well. We were able to test and verify their performance. But next comes the hard part: putting it all together to form a functional unit. Due to time constraints we did not manage to test the complete system. Nonetheless, there are still areas of improvement that pose a possibility for future research. One of which is the calibration of the delicate balance between the toxin and antitoxin concentration. Cells do an exceptional job at regulating the quantity of these molecules, which has to be mimicked for a practical autoregulatory system. Assumably, the promoters have to be adjusted or modified to reproduce the proportions of toxin to antitoxin in a stable manner. We did not implement the antitoxin in our experiments yet since this wasn’t in the realm of our experiments/since we assessed pure toxin experiments and a focus on other parts of the project more feasible and promising. Another part of the project that demands more attention are the optimal conditions to further boost AdoCbl production. Ideal synthesis circumstances have to be identified to provoke an increased yield. This would also include testing how this BioBrick compares to established production methods after completed calibration. The problem every autoregulatory system eventually has to face is the fact that even after the desired production stops for whatever reason, the synthesized molecule is still present in the cells and able to trigger the riboswitch. Since this BioBrick is regulated by recognizing the product, there is no way for it to distinguish molecules left over from the last time the cells were able to perform the relevant reaction from those that were “freshly produced”. This can result in a lack of toxin expression even if the selected compound is no longer formed. Only after the compound is successfully broken down or consumed by the cells, the intended regulation can take place. Depending on the stability of the metabolite chosen in the final system, degradation times can vary.
References
[1] Hallberg ZF, Su Y, Kitto RZ, Hammond MC. Engineering and In Vivo Applications of Riboswitches. Annual Review of Biochemistry [Internet]. 2017 Jun 20;86(1):515–39. Available from: https://doi.org/10.1146/annurev-biochem-060815-014628
[2] Fowler CC, Brown ED, Li Y. Using a Riboswitch Sensor to Examine Coenzyme B12 Metabolism and Transport in E. coli. Chemistry & Biology [Internet]. 2010 Jul 1;17(7):756–65. Available from: https://doi.org/10.1016/j.chembiol.2010.05.025
[3] Sheppard DE, Penrod JT, Bobik TA, Kofoid E, Roth JR. Evidence that a B 12 -Adenosyl Transferase Is Encoded within the Ethanolamine Operon of Salmonella enterica. Journal of Bacteriology [Internet]. 2004 Nov 15;186(22):7635–44. Available from: https://doi.org/10.1128/jb.186.22.7635-7644.2004
[4] Fa S, Jm T. Microbial Metabolism of Amino Alcohols. Ethanolamine Catabolism Mediated by Coenzyme B12-dependent Ethanolamine Ammonia-Lyase in Escherichia coli and Klebsiella aerogenes. Journal of General Microbiology [Internet]. 1976 Jul 1;95(1):173–6. Available from: https://doi.org/10.1099/00221287-95-1-173
[5] Dong L, Fang H, Gai Y, Zhao J, Jiang P, Lei W, et al. Metabolic engineering and optimization of the fermentation medium for vitamin B12 production in Escherichia coli. Bioprocess and Biosystems Engineering [Internet]. 2020 May 12;43(10):1735–45. Available from: https://doi.org/10.1007/s00449-020-02355-z
[6] Nou X, Kadner RJ. Adenosylcobalamin inhibits ribosome binding to btuB RNA. Proceedings of the National Academy of Sciences of the United States of America [Internet]. 2000 Jun 13;97(13):7190–5. Available from: https://doi.org/10.1073/pnas.130013897
[7] Borujeni AE, Mishler DM, Wang J, Huso W, Salis HM. Automated physics-based design of synthetic riboswitches from diverse RNA aptamers. Nucleic Acids Research [Internet]. 2015 Nov 30;44(1):1–13. Available from: https://doi.org/10.1093/nar/gkv1289
[8] Brzozowska I, Zielenkiewicz U. Regulation of toxin–antitoxin systems by proteolysis. Plasmid [Internet]. 2013 Jul 1;70(1):33–41. Available from: https://doi.org/10.1016/j.plasmid.2013.01.007
[9] Marianovsky I, Aizenman E, Engelberg‐Kulka H, Glaser G. The Regulation of the Escherichia coli mazEF Promoter Involves an Unusual Alternating Palindrome. Journal of Biological Chemistry [Internet]. 2001 Feb 1;276(8):5975–84. Available from: https://doi.org/10.1074/jbc.m008832200
Sequence and Features
- 10COMPATIBLE WITH RFC[10]
- 12COMPATIBLE WITH RFC[12]
- 21INCOMPATIBLE WITH RFC[21]Illegal BglII site found at 710
Illegal BamHI site found at 2756 - 23COMPATIBLE WITH RFC[23]
- 25INCOMPATIBLE WITH RFC[25]Illegal NgoMIV site found at 1603
- 1000INCOMPATIBLE WITH RFC[1000]Illegal BsaI site found at 2238
Illegal BsaI site found at 2484
Illegal BsaI site found at 3045
| None |