To identify genes that are upregulated in response to various stresses in E. coli, To consult this literature, please visit the respective part pages of each promoter. A full list of our parts can be found on our parts page literature was consulted. Next, to isolate the promoters of these genes, a 300 bp region upstream of the stress-related genes was amplified by PCR from E. coli genomic DNA. Visit our primers page for a list of primers used Forward primers contained EcoRI, XbaI overhangs and reverse primers contained PstI and SpeI overhangs, in order to create parts in the BioBrick format (Figure 1). Visit our primers page for a list of primers used. The promoter fragments were subsequently ligated into chloramphenicol backbones (pSB1C3). See figure 2 for an overview of all promoters and their stress pathway of origin. To ensure future teams and other researchers will be able to use our work, we submitted a total of 26 stress-related promoters and one novel constitutive promoter (see the chromoprotein section) as Visit our Parts to see all the parts we submitted to the iGEM registry basic parts to the iGEM part registry.

For the creation of reporter strains, our isolated promoters were cloned upstream of GFP, mCherry, AmilCP (blue chromoprotein) or spisPink (pink chromoprotein) reporter genes. Cloning was performed using the Visit our Protocols page for more information BioBrick assembly method. Correct insertion was verified by restriction analysis with NdeI and PvuI, see figure 3. 18 of these constructs were submitted as Visit our Parts page to view all the parts we submitted to the iGEM registry composite parts.

Conclusion

26 promoter regions of stress-related genes and one novel constitutive promoter were successfully amplified from the E. coli genome, ligated into the chloramphenicol backbone (pSB1C3) and submitted as BioBrick basic parts. These promoters were used to create reporter strains with GFP, mCherry, AmilCP and spisPink as reporter genes, 18 of which were submitted as BioBrick Visit our Parts to see all the parts we submitted to the iGEM registry composite parts. �

Characterising our promoter repertoire

The goal of our system is to enable detection of specific types of cellular stress using reporter strains. To this end, we performed characterisation experiments of our GFP reporter strains using flow cytometry. To establish which reporter strains could successfully detect different types of cellular stresses, we incubated cells with a range of stressors with different targets in the cell, see table 1 for an overview.

Table 1. Stress pathways targeted and stressors used in the initial characterisation of our GFP reporter strains.

Stress pathway	Stressor
Oxidative stress	Hydrogen peroxide
Envelope stress	Ethanol
DNA damage	Nalidixic acid, zeocin
RNA synthesis	Rifampicin
Protein synthesis	Apramycin, chloramphenicol, hygromycin, kanamycin, streptomycin, tetracycline
Cell wall synthesis	Ampicillin, carbenicillin, vancomycin

After measuring using flow cytometry, results of this preliminary screening were summarised in a heatmap showing relative expression levels which were normalised to incubation without stressors (Figure 4). For a detailed description of the measurement process, see our Protocols and Measurement pages.

Our heatmap shows large variation in the response of promoters to our library of stressors, thereby showcasing the potential of a system utilizing stress promoters to determine and distinguish bacterial stress types. Of the 24 promoters we tested, 8 promoters showed significant upregulation in response to one or a few stressors from our utilized library, shown in table 2. Promoters with no significant response will need further validation using additional stressors to fully exploit their use.

Reporter construct	BioBrick code	Inducing stressor
pKatG-GFP	BBa_K2610018	H2O2
pRelA-GFP	N/A	Carbenicillin, streptomycin
pSoxS-GFP	BBa_K2610031	H2O2, nalidixic acid
pSecA-GFP	BBa_K2610011 (promoter only)	Kanamycin
pRpoE-GFP	BBa_K2610021 (promoter only)	Ethanol, chloramphenicol, nalidixic acid
pSpy-GFP	BBa_K2610025 (promoter only)	Ethanol, chloramphenicol, nalidixic acid
pInaA-GFP	BBa_K2610026 (promoter only)	H2O2, nalidixic acid
pClpB-GFP	BBa_K2610020	Ethanol

Validating our promoters using confocal microscopy

The pSoxS-GFP (BBa_K2610031) BioBrick was also validated using confocal microscopy. Incubation of our reporter strain with nalidixic acid can be seen to increase GFP expression compared to basal GFP expression levels (Figure 5).

One of our BioBricks, pCspA-GFP (BBa_K2610037), showed a significant downregulation as a specific response to all protein synthesis inhibitors in our library. It should be noted that this response could be caused by inhibition of GFP production due to the protein synthesis stressors. Similarly, in our Measurement page we show that protein synthesis inhibitors lead to lower observed GFP expression for constitutive promoters, the same could be occurring with pCspA which has a high basal expression level.
However, we still believe this result to be relevant as other promoters show no apparent specific downregulation by protein synthesis inhibitors only. Furthermore, we validated this construct further using confocal microscopy (Figure 6). We induced the pCspA-GFP (BBa_K2610037) BioBrick with kanamycin and included a pSoxS-mCherry (BBa_K2610032) as a positive control for protein synthesis. Although the base-level signal of mCherry is much lower than that of GFP, we can see that GFP expression is selectively inhibited when the strains are treated with kanamycin while mCherry is not.
Whether the culprit of decreased GFP expression is protein synthesis inhibition or promoter regulation could be confirmed further in the future by using a Tet-On Tet-Off system^[3]. Here, downregulation of pCspA would cause an increase in GFP expression. A response of pCspA due to protein synthesis inhibitors could then be confirmed by the observation that incubation with such agents would lead to an increase in GFP production.

Determining promoter dose dependency

We validated the GFP reporters pKatG (BBa_K2610018), pSoxS (BBa_K2610031), pSpy (BBa_K2610025), pRpoE (BBa_K2610021) and pInaA (BBa_K2610026) further with a dose-dependence study using the stressors they were originally found to respond strongest to, the results of which are shown in figure 7. For all promoters, we see a clear dose-dependent response. Our pSoxS-GFP (BBa_K2610031) BioBrick in particular displays the power of our system. An increase of GFP expression can be observed at a concentration 20 times lower than the observed lethal dose.

Conclusion

Flow cytometry measurements with promoter-GFP constructs show that reporter strains for different types of cellular stress can be selectively and dose-dependently induced by different types of stressors. The selectivity in particular establishes an important proof of concept for our system, as this shows that multiple reporters could theoretically be combined to detect different stresses in a single cell line. Of 26 promoters, 8 showed a significant response to our library of stressors. More research is needed to discover inducing compounds for the remaining reporter strains.

We have established that our stress reporters can be used to detect cell stress at stressor concentrations much lower than the lethal dose. To improve these results, we attempted to increase the dose response GFP expression even further, to allow for detection of even lower concentrations of stressful substances. Additionally, increased signal intensity would facilitate GFP-based stress detection with less sensitive detection methods. Furthermore, a readout may be obtained in a shorter time period, which would increase throughput of compound screening. This amplification method could also be applied to our visual reporters based on chromoproteins (discussed further in the visual stress detection system section). Although we have shown our chromoprotein stress reporters work, the readout is not easily visible and only occurred after two to three days.

We attempted to amplify the reporter signal in two ways. First of all, we tried to create a T7 RNA amplification system. For this, we coupled stress promoters to the T7 RNA Polymerase gene (BBa_K2610007). This should result in the T7 RNA Polymerase being produced in response to stress. Next, we coupled a T7 Promoter (BBa_K2610005) - which can only be read by the T7 RNA Polymerase - to visual reporters. Unfortunately, we were not able to validate a working T7 enhancement system. Despite this, we have submitted our pSoxS-T7 RNA Polymerase (BBa_K2610028) and pT7-GFP (BBa_K2610029) BioBricks to the registry in order to give others the opportunity to test this system.

Thereafter, we tried increasing the amount of promoter regions and GFP genes on one plasmid, see table 3 for the constructs used.

Table 3. Constructs used for signal amplification

Construct	BioBrick code
pSoxS-GFP	BBa_K2610031
pSoxS-GFP-pSoxS-GFP	BBa_K2610034
pSoxS-GFP-GFP-pSoxS-GFP-GFP	BBa_K2610035

Using these amplified BioBricks, a nalidixic acid dose-dependency study was performed. A clear increase in both absolute and relative signals compared to the negative controls can be observed in the amplified BioBricks (pSoxS-GFP-pSoxS-GFP and pSoxS-GFP-GFPpSoxS-GFP-GFP), compared to pSoxS-GFP (Figure 8). This shows that the GFP signal can be amplified by simple multiplication of promoter regions and GFP genes on one plasmid.

A disadvantage of this multiplication is that it also increases the basal expression level. Nevertheless, we observe an increase in relative expression for the amplified BioBricks, meaning they are still beneficial. In the future, we hope to increase signal strength further and keep the GFP expression as low as possible when the strain is not stressed. Further research into the T7-RNA-polymerase amplification system could provide the answer to this challenge.

Conclusion

We successfully amplified the pSoxS-GFP stress reporter strain by increasing the amount of promoters and GFP genes in the plasmid. This enables detection of even lower concentrations of stressful substances, allows signals to be detected by less sensitive detection devices and facilitates faster signal detection. When a visual reporter such as a chromoprotein is used, the time to obtain a significant readout may also be reduced.

With earlier experiments, we have shown that our system can be used to detect lethal substances in low - stressful - concentrations. In our final product, we want to be able to detect non-lethal stressful substances for use in synergistic combination therapies with other stressful substances. In this chapter, we show that our system can be used to detect stress caused by non-lethal compounds which are capable of enhancing the effect of antibiotics.

Verification of non-lethality

To be able to test non-lethal stressful substances, we first had to ensure substances we were using were not lethal. To that end, we tested 19 natural compounds, or extracts thereof, available in the supermarket and sometimes heralded in popular culture for their medicinal properties. We first assessed the growth of E. coli following addition of the potential stressors compared to the addition of water. Three of these compounds were found to be lethal in the concentration used, as seen in figure 9. Due to time constraints, we decided to test three compounds which did not shows lethality and growth inhibitions, and we thought were interesting based on literature. These three substances are ascorbic acid^[4][5], cannabis^[6][7] and ginger^[8][9].

Stress induction of non-lethal compounds

Extracts of ginger and cannabis and a vitamin C solution were incubated with GFP reporter strains containing promoters of the following genes: dnaK, cutF, clpB, rseA, cspA, relA, and nusA, results of which are shown in figure 10. These promoters were chosen for two reasons: 1) they originate from a broad spectrum of stress pathways and 2) we wanted to validate promoters that did not show a significant response in the preliminary screening (with the exception of clpB).

The highest response was seen for ascorbic acid with the clpB promoter. We therefore decided to continue with ascorbic acid and performed a dose-dependence study. To exclude the possibility of stress occurring due to the acidity of ascorbic acid, we included pH controls. These were determined by measuring the pH of ascorbic acid solutions in LB and bringing fresh LB samples to the same pH using concentrated HCl. The results are shown in figure 11.

Ascorbic acid is seen to induce significant stress not observed in the pH control for the clpB promoter at a concentration of 2.5 mg/mL. However, the GFP response decreases sharply at 5 mg/mL and rises again at 10 mg/mL. This is contrary to expectations, as we expected to see a dose response, similar to figure 7. Unfortunately, due to time constraints we were unable to repeat the experiment to rule out an experimental error. Still, based on the results in figures 9 and 10 we can draw the cautious conclusion that ascorbic acid induces the clpB promoter and is stressful at these concentrations.

Having confirmed that ascorbic acid induces cellular stress, we set out to confirm its synergistic use in combination with antibiotics. The MIC values of the ampicillin, apramycin, chloramphenicol, nalidixic acid, tetracycline and streptomycin were determined with and without the addition of ascorbic acid (Figure 12). A clear drop in the MIC value was observed for ampicillin, chloramphenicol and nalidixic acid, up to 82% in the case of chloramphenicol. This is a remarkable result considering these three antibiotics all have different modes of action and ascorbic acid is apparently able to enhance all of them.

Conclusion

In the future, we intend our stress detection system to be used for discovery of non-lethal compounds that induce cellular stress. These substances can synergistically enhance current antibiotics or be used in novel combination therapies. We have shown that ascorbic acid is an ideal candidate compound. First, we verified its non-lethality. Next, we observed that it induced the clpB promoter showing that it is a stressful compound. Finally, we proved that ascorbic acid enhances the effect of ampicillin, chloramphenicol and nalidixic acid. These experiments demonstrate that our system is capable of detecting stressful compounds, which can be used in synergistic combination therapies.

A new positive control - Isolation and characterisation of a novel constitutive promoter

During our project, we isolated the GapA promoter (BBa_K2610040) from the E. coli genome. Contrary to our stress promoters, this promoter has a constitutive high basal expression level, as can be seen in the spun down cultures of our pink and blue chromoprotein BioBricks: pGapA-AmilCP (BBa_K2610041) and pGapA-spisPink (BBa_K2610042), Figure 13. In contrast to many synthetic promoters, this novel promoter provides an indication of high levels of natural expression. Therefore, this promoter can be used as a future positive control of reporter production - such as GFP - in a natural situation.

Chromoprotein-based stress detection for overlays and disk diffusion

Experiments in which potential antibiotic producing strains or disks containing antibiotics are placed on a plate, after which a bacterial test strain is seeded on top. Compounds diffuse through the agar and - if lethal - prevent bacterial growth surrounding the strain or disk, thereby creating a visible “zone of inhibition" Overlays and disk diffusions have been pivotal in the past for the discovery and characterisation of antibiotics^[10][11]. Because these experiments are carried out on agar plates, a GFP-based readout is not suitable, since GFP production would be cumbersome to measure reproducibly on a plate. Furthermore, quantitative measurements are not essential in this case, as these techniques are purely used to identify interesting compounds or antibiotic producing strains. Therefore, we chose for a Read more about this decision on our Integrated Human Practices page readout visible to the naked eye in the form of chromoproteins. Disk diffusion and overlay experiments display halos where bacteria cannot grow. At the edge of such a halo, there is a zone with a Read more about this zone, and the determination of it on our Model page non-lethal antibiotic concentration. In this zone we could potentially detect cellular stress. To test this, we created a plasmid containing a blue chromoprotein regulated by a stress promoter, pSoxS-AmilCP (BBa_K2610033) and pGapA-spisPink (BBa_K2610033). From our flow cytometry experiments, we knew that pSoxS should react to nalidixic acid. We attempted a disk diffusion with our pSoxS-AmilCP BioBricks, which is shown in figure 14. A discoloration around the halo can be seen, which is not seen in the negative control. This shows we can visualize bacterial cell stress using chromoproteins.

We also attempted to achieve similar results in liquid cultures by incubating bacteria expressing the pSoxS-AmilCP BioBrick with increasing concentrations of nalidixic acid, results of which are shown in figure 15. However, we found no discernible difference in color for the stressed vs. unstressed strains. Evidently, the stress caused by nalidixic acid at our used concentrations is not enough to create a difference in expression visible to the naked eye. Further optimisation with, among other things, different incubation times is necessary to achieve the desired results.

Conclusion

We isolated a novel constitutive promoter from the E. coli genome and showed a high natural expression level of chromoproteins. Additionally, we have shown that chromoproteins can be used as visual reporters for cell stress in conditions representative of overlay and disk diffusion experiments. This is the first time visual reporters have been used in such a way. This opens the door for many agar plate-based detection systems. In addition, visual reporters can be used in labs where expensive plate readers and flow cytometers that measure GFP expression are not available. Since our BioBricks are available on the registry, experiments using these strains can begin today!

Conclusion

The global challenge of antibiotic resistance requires novel approaches in the field of antibiotic discovery. Two major missed opportunities in antibiotic research are the detection of non-lethal stressful compounds and the early determination of a compound’s mode of action. With Fifty Shades of Stress, we tackle both issues.

To achieve this, we isolated promoters of stress-related genes from the E. coli genome using PCR and submitted 27 of them as Read about each of our parts on our Parts page basic parts to the registry. The promoters were cloned upstream of two fluorescent reporter genes, GFP and mCherry, and two visual reporter genes (chromoproteins), amilCP (blue) and spisPink (pink). 18 of these constructs were submitted as Read about each of our parts on our Parts page composite parts.

To characterize the stress responsiveness of our promoters, we subjected 24 GFP reporter strains to a library of stressors that elicit different types of stress. 8 reporters yielded a clear increase in GFP expression in response to one or more compounds in our library. Several were validated further using confocal microscopy and/or dose-response studies. Additionally, one reporter, pCspA-GFP (BBa_K2610037), was presumably selectively downregulated, although more research is needed to confirm this.

To improve signal detection, we have amplified GFP expression by increasing the number of promoters and GFP genes in one plasmid. We performed the experiment using the pSoxS (BBa_K2610030) BioBrick and the signal increases almost tenfold. These parts were also submitted as composite parts.

To prove that our system could be used to detect stressful compounds, we tested a range of natural products for their lethality. Three of these compounds were tested on different stress promoters and ascorbic acid was found to be stress-inducing. Therefore, ascorbic acid was incubated together with several antibiotics and the MIC value was found to drop drastically, up to 82% for ampicillin. Thereby proving our system is capable of detecting stressful compounds suitable for antibiotic combination therapies.

For experiments with antibiotics on agar plates - such as disk diffusions or overlays - we set out to create a visual reporter strain using chromoproteins, an approach which has not been attempted before. We performed a disk diffusion experiment using our pSoxS-AmilCP (BBa_K2610033) BioBrick as a reporter and found that a discoloration occurred in the area where the antibiotic concentration is non-lethal. This is an important proof of concept for visual reporters to be used in agar-based experiments.

Towards an application

We have shown that the foundation for Fifty Shades of Stress is solid. The next step is creating an easy-to-use application from our project. To that end, we have created a Read about our product design in detail on our Product Design wiki page product design for an easy-to-use bacterial cell stress detection kit with the name 50S.O.S.

We want to establish a proof of concept for three important aspects of the kit before full scale production can begin. First, we want GFP expression by our reporter strains to be measurable in plate readers. This would make measurement faster and open up the possibility of obtaining real-time data. Unfortunately, we experienced repeated complications when measuring in the plate reader available to us, therefore preventing us from examining this possibility further. However, quantifying GFP expression in E. coli strains using a plate reader is not new and should work^[12]. We surmise that the cause of our failed plate reader experiments may be unsuitable transparent plates or diluting the cultures too much before stressing them.

A second goal for our future application is to enable testing of multiple stress promoters in a single bacterial strain, by using fluorescent reporters with different emission wavelengths. We attempted this with the GFP and mCherry reporters using co-transformations and by combining two reporter constructs in one plasmid. However, neither technique yielded a replication of the results obtained with the single-promoter GFP or mCherry plasmids. This may be due to a lower basal expression level of mCherry or lower fluorescence intensity. Still, we submitted two combined plasmids as composite parts for other teams and researchers to use and improve: pSoxS-mCherry-pKatG-GFP (BBa_K2610043) and pSoxS-mCherry-pMicF-GFP (BBa_K2610044). Also, we do not foresee any fundamental reason why our combination plasmids should not have worked^[13], so we are confident further research could easily allow for the combinations of multiple stress reporters in a single bacterial strain.

A third important aspect of our 50S.O.S. kit is the use of freeze dried strains, to allow for safer GMO transport. This is a common technique and should work according to literature. We have performed such freeze drying experiments with our strains using More information can be found on our Protocols page glycerol. Although we were able to revive our strains when using high glycerol concentrations, reviving was no longer successful when using lower concentrations of glycerol. Therefore, we believe future attempts should use sucrose, which is more commonly used for freeze drying E. coli^[14].

Although we did not yet achieve success with these three extra aspects for our kit design, the techniques described above all have precedents in literature. That means it is simply a matter of time before they can all be integrated into a full-fledged 50S.O.S. system. For now, researchers can use each of our publicly available stress detection strains by obtaining them from the iGEM registry. This way, we are one step closer to discovering novel antibiotic therapies and solving the antibiotic resistance crisis.

Below the data for the heatmap with the standard error and the amount of replicates included is shown.