Experiments (Build) & Results (Test)
We first verifed the function of gut inflammation thiosulfate and tetrathionate sensors by sfGFP expression in E. coli Top10 and E. coli Nissle 1917. After the experimental simulation to confirm that colorful bacteria (expressing chromoproteins) are able to work even in the condition of “excrement” (actually curry), we replaced the sensors’ reporter by gfasPurple, asPink, amilCP, respectively. While thiosulfate sensor works with chromoproteins as expected, the tetrathionate sensor does’t response very well. Accordingly we changed its reporter to a dark-green molecular-protoviolaceinic acid, which successfully solved the problem. At last, we experimentally demonstrated that E. coli cells carrying our engineered sensor-reporting systems are quite promising to detect gut inflammation in real medical scenarios.
Verifying the function of ThsS/R & TtrS/R
ThsS (BBa_K2507000) and ThsR (BBa_K2507001), both codon-optimized for E. coli, are two basic parts which belong to the two-component system from the marine bacterium Shewanella halifaxensis. ThsS is the membrane-bound sensor kinase (SK) which can sense thiosulfate outside the cell, and ThsR is the DNA-binding response regulator(RR). PphsA(BBa_K2507018) is a ThsR-activated promoter which is turned on when ThsR is phosphorylated by ThsS after ThsS senses thiosulfate.
Because thiosulfate is an indicator of intestinal inflammation (Levitt et al, 1999; Jackson et al, 2012; Vitvitsky et al, 2015), the ThsS/R system can be used as a sensor for intestinal inflammation. Our experiments confirmed that the system indeed works as a thiosulfate sensor, both in Escherichia coli Top10 and E. coli Nissle 1917. By linking thsR with sfgfp (BBa_K2507008), chromoprotein genes (BBa_K2507009, BBa_K2507010, BBa_K2507011) or the violacein producing operon vioABDE (BBa_K2507012), this system can respond to thiosulfate by producing a naked-eye-visible signal (GFP, spisPink-pink chromoprotein, gfasPurple-purple chromoprotein, amilCP-blue chromoprotein, or the dark-green pigment protoviolaceinic acid).
Figure 1. Schematic diagram of the ligand-induced signaling through ThsS/R and the plasmid-borne implementation of the sensor components. ThsS/R was tested by introducing BBa_K2507004 into the pSB4K5 backbone and BBa_K2507008 into the pSB1C3 backbone. We submitted all of the parts to the iGEM registry in pSB1C3.
Characterization experiments were performed aerobically. Bacteria were cultured overnight in a 96-deep-well-plate, with 1ml LB media + antibiotics + different concentrations of inducer (thiosulfate) in each well. The conclusion is that while the system (ThsS/R) works, despite the leaky expression is rather heavy.
Figure 2. Characterization of ThsS/R system using sfGFP. 1mM, 0.1mM, 0.01mM and 0 Na2S2O3 were added as required.
Previously, Schmidl et al. have shown that thsR overexpression in the absence of the cognate SK can spontaneously activate a strong output (Schmidl et al, 2014), possibly due to RR phosphorylation by alternative sources (small molecules, non-cognate SKs), or low-affinity binding by non-phosphorylated RRs.
This reminded us to check the plasmid design. To our surprise, we found that the plasmid backbone (pSB4K5) we used for thsR expression has several mutations in the pSC101 replication origin, which means that pSB4K5 is actually a high-copy plasmid!
Due to the limited time after we found this bug, we were unable to test the ThsS/R system on a really low-copy plasmid backbone, but we will certainly do it after the 2017 iGEM Jamboree.
Next, we characterized the system under aerobic and anaerobic conditions. We measured sfGFP intensity using flow cytometry. (http://2017.igem.org/Team:SHSBNU_China/Protocol). The response curves in aerobic and anaerobic condition seem very similar in E. coli Top10, while in E. coli Nissle 1917, GFP expression levels are a bit different in aerobic and anaerobic conditions (Figure 5)
Figure 3. Characterization of ThsS/R system in E. coli Top10 and E. coli Nissle 1917 under different conditions, by measuring the sfGFP expression levels via flow cytometry.
Figure 4. Detailed characterization of ThsS/R system using flow cytometry.
Figure 5. E. coli Nissle 1917 haboring ThsS/R system was cultivated overnight under aerobic and anaerobic conditions, respectively.
Codon-optimized TtrS(BBa_K2507002) and TtrR (BBa_K2507003) are two basic parts which are derived from the two-component system of marine bacterium Shewanella baltica. TtrS is the membrane-bound sensor kinase (SK) which can sense tetrathionate outside the cell, and TtrR is the DNA-binding response regulator (RR). PttrB185-269 (BBa_K2507019) is a minimal TtrR-activated promoter which is activated when TtrR is phosphorylated by TtrS after TtrS senses tetrathionate.
Winter et al. have shown that reactive oxygen species (ROS) produced by the host during inflammation convert thiosulfate into tetrathionate, which this pathogen consumes to establish a beachhead for infection (Winter et al, 2010). Thus, tetrathionate may correlate with pro-inflammatory conditions and can therefore be used as a sensor for gut inflammation.
We first validated that this system can function as a tetrathionate sensor and reporter in the laboratory strains Escherichia coli Top10 and E. coli Nissle 1917.
Figure 6. Schematic diagram of ligand-induced signaling through TtrS/R and plasmid-borne implementation of the sensor components. TtrS/R was tested with BBa_K2507006 integrated into the pSB4K5 backbone and BBa_K2507013 into the pSB1C3 backbone. We submitted all parts to the iGEM registry in pSB1C3.
Figure 7. Tetrathionate sesnor system does not work well initially.
Figure 8. E. coli Nissle 1917 haboring TtrS/R system was cultivated overnight under aerobic or anaerobic condition.
Selection of an appropriate chromoprotein
Figure 9. GFP maturation time course of anaerobically grown E. coli Nissle 1917 in PBS+1 mg/mL chloramphenicol, adapted from Kristina et al., 2017. sfGFP fluorescence increases over time for ThsS/R in the presence and absence of saturating thiosulfate.
Human excrements have colors ranging from yellow to brown and even black. Thus, the bacteria in excrement must show a contrasting color in order to efficiently indicate the presence of gut inflammation.
At the beginning, we tested six different chromoproteins to find the most suitable candidate which will possibly be used as the reporter. We used pSEVA321(From Bluepha) as the backbone with a strong constitutive promoter. The following proteins were tested:
1. BBa_K1033910 "fwYellow yellow chromoprotein"
2. BBa_K1033916 "amajLime yellow-green chromoprotein"
3. BBa_K592010 "amilGFP yellow chromoprotein"
4. BBa_K1033919 "gfasPurple purple chromoprotein"
5. BBa_K1033932 "spisPink pink chromoprotein"
6. BBa_K592009 "amilCP blue chromoprotein”
Figure 10. Schematic diagram of chromoprotein reporters.
After creating bacteria with different colors, we decided to conduct a color test with these bacteria to show which bacteria can best show the color.
First, we grew cultures of bacteria expressing each of the chromoprotein constructs we designed,
Figure 11. Overnight cultivated E. coli haboring different chromoprotein expression plasmids.
After centrifugation, we obtained the following bacterial pellets:
Figure 12. Centrifuged Bacterial pellets.
Subsequently, we used curry to imitate the color of excrement, and mixed the bacteria with the resulting paste, obtaining the following results:
Figure 13. Mix centrifuged bacteria with 1g curry and 500uL water.
Pink, blue, and purple chromoproteins are the most distinguishable. Therefore, we decided to use them as reporters towards real application scenarios.
Replacing sfGFP with chromoproteins
After selecting three chromoproteins, we used Golden Gate method to replace gfp gene with chromoprotein genes. We successfully constructed BBa_K2507009, K2507010, K2507011, K2507014, K2507015, K2507016 and Co-transformed with BBa_ K2507004/ BBa_K2507006 into E. coli Top10.
Figure 14. Schematic diagram of the ligand-induced signaling through ThsS/R and the plasmid-borne implementation of the sensor components. We combine BBa_ K2507004 with BBa_K2507009, BBa_K2507010, BBa_K2507011. b. Schematic diagram of ligand-induced signaling through TtrS/R and plasmid-borne implementation of the sensor components. We combine BBa_ K2507006 with BBa_K2507014, BBa_K2507015, BBa_K2507016.
We cultivated the bacteria in different concentration of thiosulfate or tetrathionate. After validating the system in the laboratory strains E. coli Top10, we confirmed that the system with chromoproteins indeed works as a thiosulfate sensor, as intended. By linking thsR with chromoprotein genes (BBa_K2507009, BBa_K2507010, BBa_K2507011) , this system can respond to thiosulfate by producing a signal visible to the naked eye.
Figure 15. ThsS/R-chromoprotein systems work as expected, while TtrS/R-chromoprotein systems don’t work.
Replacing sfGFP with vioAVBDE pathway
Then we found that Cambrigde iGEM 2009 developed and characterized several pigments. So we want to use violacien to be the reporter. After searching former iGEM part distribution kit in Beijing, finally we get the plasmid BBa_K274003 from Peking iGEM team a 2010 iGEM part distribution kit. BBa_K274003 coding vioABDE which coding enzymes produce a precursor of violacien- protoviolaceinic acid, which is dark-green.
Figure 16. The violacein biosynthetic pathway (From http://parts.igem.org/Part:BBa_K274002). Genes for violacein biosynthesis are arranged in an operon consisting of vioA, vioB, vioC, vioDand vioE. VioA generates an IPA imine from L-tryptophan and VioB converts the IPA imine into a dimer. VioE then acts by transforming the dimer into protodexyviolaceinic acid (PVA), which can be spontaneously converted into a green pigment called deoxychromoviridans. VioD and VioC hydroxylate PVA to form violacein.
Figure 17. Schematic diagram of the ligand-induced signaling through ThsS/R and the plasmid-borne implementation of the sensor components. We combine BBa_ K2507004 with BBa_K2507012. b. Schematic diagram of ligand-induced signaling through TtrS/R and plasmid-borne implementation of the sensor components. We combine BBa_ K2507006 with BBa_K2507017.
Figure 18. TtrS/R system works better than ThsS/R system when combined with protoviolaceinic acid biosynthesis pathway, producing naked-eye-visible output.
On November 16, 2011, Pharma Zentrale company sent an application to FDA and states that each their capsule contains E.coli strain 1917 corresponding to 2.5-25×109 viable cells.
We assumed 2.5×109 bacteria per capsule. The total weight of one cell is 9.5×10-13g. Therefore, the weight of bacteria in one capsule is about 0.024g.
Figure 19. Bacterial dose in simulated application scenarios. a. Engineered bacteria were cultured overnight and then centrifuged. b. 1g curry, 500ul H2O were added, resuspended and centrifuged. c. 2g curry, 500ul H2O were added, resuspended and centrifuged. 1 and 2, 50ml LB+ TtrS/R- vio system in E.coli Top10. 3 and 4, 50ml LB+ ThsS/R- Pink system in E.coli Top10. Pink, spisPink; vio, protoviolaceinic acid.
Sample each of ttr-vio (ttrS/R sensor + protoviolaceinic acid biosynthesis pathway) and ths-pink (thsS/R sensor + spisPink chromoprotein) overnight cell culture for 100 uL and plate them on two different plates, we got:
Therefore, there are 5.2×109 bacteria in 20mL ttr-vio cell culture, corresponding to a weight of ~0.050g.There are 7.6×109 bacterial cells in 20mL ths-pink cell culture, corresponding to a weight of ~0.072g; for 50mL cell culture, the weight is ~0.180g.
Daeffler, K. N., Galley, J. D., Sheth, R. U., Ortiz‐Velez, L. C., Bibb, C. O., & Shroyer, N. F., et al. (2017). Engineering bacterial thiosulfate and tetrathionate sensors for detecting gut inflammation. Molecular Systems Biology, 13(4), 923.
Frederick C. Neihardt (1996), Escherichia coli and Salmonella: Cellular and Molecular Biology (1st volume), ASM Press. Available at: http://kirschner.med.harvard.edu/files/bionumbers/Composition%20of%20an%20average%
Jackson MR, Melideo SL, Jorns MS (2012) Human sulfide: quinone oxidoreductase catalyzes the first step in hydrogen sulfide metabolism and produces a sulfane sulfur metabolite. Biochemistry 51: 6804 – 6815
Levitt MD, Furne J, Springfield J, Suarez F, DeMaster E (1999) Detoxification of hydrogen sulfide and methanethiol in the cecal mucosa. J Clin Invest 104: 1107 – 1114
Schmidl SR, Sheth RU, Wu A, Tabor JJ (2014) Refactoring and optimization of light-switchable Escherichia coli two-component systems. ACS Synth Biol 3: 820 – 831
Vitvitsky V, Yadav PK, Kurthen A, Banerjee R (2015) Sulfide oxidation by a noncanonical pathway in red blood cells generates thiosulfate and polysulfides. J Biol Chem 290: 8310 – 8320