GENETIC CIRCUIT

Our genetic circuit consists of several parts, the most important of which include mphR, a transcription factor that acts as a repressor in the absence of macrolides [1], ermC, a macrolide antibiotic resistance gene [2], and egfp, which codifies for the green fluorescent protein [3], controlled by a promoter regulated by MphR (pMphR) [1]. The system works in the following way (Fig. 1): when there are no macrolides present in the sample, MphR binds to pMphR preventing the expression of the egfp gene. Therefore, if there are no macrolides, the biosensor does not give any fluorescence output. On the other hand, if there are macrolides in the sample, these bind to MphR, and the whole complex is unable to bind to the DNA. In this case, there is an expression of the egfp so the user can see a fluorescence output. Noteworthy mentioning, ermC, which is regulated by the same promoter as mphR, gives our biosensor resistance to macrolides by methylating the rRNA.

MACROLIDES' MODE OF ACTION

Macrolide antibiotics are bacteriostatic, meaning they prevent bacterial growth. They achieve that by binding to the 23S rRNA subunit, thus preventing peptide bond formation. This hinders the cell’s translational machinery and inhibits any protein synthesis. An additional proposed mechanism might be triggering dissociation of the peptidyl-tRNA during its translocation within the ribosome [4].

REASONING FOR THE BIOSENSOR DESIGN

The first decision in the biosensor design is choosing between whole cell and molecular biosensors, which are usually based on enzymes on antibodies outside of the living organism. Whole cell biosensors have several advantages. Firstly, they are cheaper to produce, since there is no need for protein purification and cells are self-replicating. They are also easier to both engineer and use, they do not require complex calibration protocols, which makes them an interesting alternative for laboratories or research projects without enough budget for chemical or molecular detection methods. Another important advantage is the fact that the devices can have relatively small sizes [5]. Lastly, whole cell biosensors show tolerance for a greater range of environmental conditions [6]. For these reasons, our team decided to focus on developing a whole cell biosensor instead of a molecular biosensor. Yet, we still had to keep in mind potential issues, such as a background signal, which may negatively impact the detection limits. Proliferation of cells, even though it is seemingly an advantage, also may render outcome unreliable and should be properly controlled, for example through immobilization [6].

The next crucial step is choosing a microorganism. We have chosen Escherichia coli. This was the most logical choice since it is a model microorganism very well characterized. Regarding the choice of the strain, we have worked with two different strains: BL21 and TOP10. BL21 has an improved protein synthesis machinery and has been previously used in environmental biosensors [6]. We chose this strain since we wanted our biosensor to produce enough green fluorescent protein. TOP10 cells were mostly used for our transformation experiments, since we observed higher transformation efficiency.

To ensure a sufficient output signal, we used a constitutive promoter to regulate the mphR and ermC expression. After obtaining initial results, we also wanted to test our device with mphR and ermC under the regulation of an inducible promoter and compare it with the constitutive promoter. The reason for it was the risk that a high amount of MphR may have continued to repress the pMphR promoter, since there would always be some of the MphR left unbound by macrolides. In these experiments, we used E. coli TOP10 strain since the inducible promoter we chose, pBAD, is not compatible with BL21, since they are not araBADC negative and araEFGH positive [7].

Since our biosensor is aimed towards macrolide detection, it is crucial that the bacteria are resistant to them. It is also important that the mechanism of resistance does not pump macrolides out of the cells or modify their structure, since this would corrupt our ability to sense these compounds. For that reason we chose ermC as the resistance gene, since its mechanism is based on ribosome methylation [8].

Another important aspect of a biosensor is its output signal. It should be easy to measure and quantify. Initially, we were considering an electrochemical biosensor, since it seems to be more sensitive [9], which is an important feature given the very low concentration of macrolides. Moreover, we are aware that with an electrochemical biosensor it would be easier to quantify the output. However, for a proof of concept, an optical biosensor would be more suitable, since it is less challenging to create and test. Due to this fact, we explored the use of Mtr pathway in E. coli just in theory, more details about which can be found here. We have researched which of viable reporter genes suits our project (Table 1). We have chosen GFP, since it is widely used in iGEM, does not require addition of substrate, it has high stability and the signal can be strengthened by increasing the amount of excitation light [10].

**Table 1.** Summary of researched reporter genes.
Gene	Protein	Does it require substrate?	Measurement method	Other
luc	Luciferase	Yes	Bioluminescence	Not appropriate for flow cytometry analysis [11]
gfp	Green Fluorescent Protein	No	Fluorescence	Increasing the amount of excitation light [12]
lacZ	β-galactosidase	Yes	Enzymatic	High levels of expression affect cell growth [13]

Since wastewater is far from the optimal environment for E. coli growth, we needed to take into account potential interfering factors and the fact that macrolide concentrations can vary greatly [14], there have been concerns that the bacteria will not have environmental pressure to retain the plasmid. An additional issue may be the presence of inhibitors, such as heavy metals, in wastewater, which would hinder GFP expression.

CONTROL OF CELL REPLICATION

Since bacteria are living organisms, it is important to control their growth to increase output signal’s stability and reliability. This can be done by cell immobilization. Properly chosen method could also increase the shelf-life of our biosensor and ensure unified access of cells to media, wastewater and macrolides. It was important that the immobilization method is cost-effective, non-toxic and biocompatible. A great resource turned out to be one of previous iGEM projects: Team Sydney has done a very concise mini-review of some immobilization methods suitable for use in whole cell biosensors. Based on our research the most promising options were alginate, latex, porous silica and poly-lysine [15, 16]. Due to limited availability of lab spaces we were only able to perform experiments with poly-lysine and alginate, the results of which can be found here. Based on our mentors’ advice we also performed tests with egg white and Cultrex.

PLANNED EXPERIMENTS

To ensure the cells chosen for biosensor are suitable, we conducted a plate reader experiment, where the cells were grown in LB as positive control, milliQ as negative control, wastewater and mixes of LB and wastewater in different proportions.
As a method to construct the genetic circuit, we have chosen Modular Cloning, since it is a golden gate cloning technique, thus is compatible with iGEM standards. In addition to that, it is modular, which makes it easier to determine which elements of the circuit are malfunctioning and allows us to replace various elements. Based on our mathematical model and results we can then select promoters or ribosomal binding sites with the most optimal strengths.
Estimating the range of the biosensor with mathematical modelling.
To determine the functionality of the ermC gene, as well as its specificity to certain macrolide antibiotics, the cells were grown in LB containing various amounts of erythromycin, clarithromycin and spectinomycin.
To see if the complete genetic circuit is functional and that the amount of GFP correlates to the amount of antibiotics, as well as to determine the detection limits of the biosensor, we conducted another plate reader experiment, where the cells were grown in media containing various amounts of erythromycin.
Assessing the potential of various immobilization methods was achieved by performing live-imaging.
To tackle the anticipated sensitivity issue, we performed a proof-of-concept isoelectric focusing experiments. We also modelled MphR mutants with potentially increased macrolide binding affinity and FhuA membrane protein with more open confirmation.

More details concerning the experiments can be found at our experiments, model, results and engineering pages.

IMPROVED DETECTION OF MACROLIDES

The most challenging aspect of macrolide antibiotics quantification is their low concentrations in wastewater. Our dry-lab’s model has confirmed that in order to achieve satisfactory detection limits and reliable output signal, concentration of macrolides may be necessary. We have explored several available methods: microfluidics and modified E. coli strains with disabled efflux pumps and modified pore proteins.

Microfluidics

Isoelectric focusing is able to concentrate an analyte up to a thousand times [17]. This method allows us to increase the macrolide concentration from the wastewater to more easily detectable levels. Paper based isoelectric focusing microfluidics are cheap and easy to produce and distribute, thus very suitable for pre-treatment of the sample, given they also work with small molecules and not only larger proteins. [18] As a proof of concept we performed tests using isoelectric focusing on gels. More details can be found on our experiments page.

Pumps and Transporters

GKCW104 E. coli strain created by Krishnamoorthy and colleagues (2016) has deleted tolC genes. TolC are proteins that form an outer membrane channel in several multidrug resistance efflux pumps. Moreover, there were additional mutations introduced, which permanently open FhuA proteins, leading to hyperporation of the cell [19].

Membrane Permeability

We were also considering using chemicals with permeabilizer properties, such as lactic acid. Disrupting will most likely affect GFP production and overall function of the cell, but it may be possible to find a balance on disrupting the cell wall enough so that functions of the cells are not changed, yet more erythromycin can enter the cell. Due to limited time in the lab we did not manage to experimentally test this approach.

More about our concentration research can be found here.

BETTER SENSITIVITY - MphR MUTANTS

Also, a different way of improving biosensor sensitivity included selection of MphR mutants with alterations at the binding site that would increase the binding affinity of MphR for macrolides and thus, potentially lower the detection limit. We modelled the MphR binding sites with Rosetta and from the results, selected the two most promising mutants with higher binding affinity for clarithromycin and three for erythromycin. We also selected a mutant with high binding affinity for both erythromycin and clarithromycin. We ordered the 5 most promising sequences to test experimentally in the lab. The results of the above mentioned modelling can be found here.

BIOSENSOR SUMMARY

Plasmid: Copy Number

High Copy Number.
Low Copy Number: Remains to be tested experimentally.

Escherichia coli BL21

Increased protein production.

Immobilization

Requires further research.

Transcription Factor: MphR

High specificity for macrolide antibiotics.

Macrolide Resistance: ermC

No modification of macrolides.
Acts through methylation of rRNA.

Optical Output: EGFP

Commonly used in iGEM competition.
Does not require substrates.

REFERENCES

1. Zheng, J., Sagar, V., Smolinsky, A., Bourke, C., LaRonde-LeBlanc, N., & Cropp, T. A. (2009). Structure and function of the macrolide biosensor protein, MphR(A), with and without erythromycin. Journal of Molecular Biology, 387(5), 1250-1260. doi:10.1016/j.jmb.2009.02.058
2. UniProt: a worldwide hub of protein knowledge. (2018). Nucleic Acids Research, 47(D1), D506-D515. doi: 10.1093/nar/gky1049
3. Cormack, B., Valdivia, R., & Falkow, S. (1996). FACS-optimized mutants of the green fluorescent protein (GFP). Gene, 173(1), 33-38. doi: 10.1016/0378-1119(95)00685-0
4. Kanoh, S., & Rubin, B. K. (2010). Mechanisms of action and clinical application of macrolides as immunomodulatory medications. Clinical Microbiology Reviews, 23(3), 590-615. doi:10.1128/cmr.00078-09
5. Harms, H., Wells, M., & van der Meer, J. (2006). Whole-cell living biosensors—are they ready for environmental application?. Applied Microbiology And Biotechnology, 70(3), 273-280. doi: 10.1007/s00253-006-0319-4
6. Xu, X., & Ying, Y. (2011). Microbial biosensors for environmental monitoring and food analysis. Food Reviews International, 27(3), 300-329. doi:10.1080/87559129.2011.563393
7. https://www.thermofisher.com/fi/en/home/ technical-resources/technical-reference-library/protein-expression-support-center/bacterial-expression-support/bacterial-expression-support-getting-started.html
8. Dinos, G. P. (2017). The macrolide antibiotic renaissance. British Journal of Pharmacology, 174(18), 2967-2983. doi:10.1111/bph.13936
9. Yi, H., Li, M., Huo, X., Zeng, G., Lai, C., & Huang, D. et al. (2019). Recent development of advanced biotechnology for wastewater treatment. Critical Reviews In Biotechnology, 40(1), 99-118. doi: 10.1080/07388551.2019.1682964
10. Kain, S. R., Adams, M., Kondepudi, A., Yang, T. T., Ward, W. W., & Kitts, P. (1995). Green fluorescent protein as a reporter of gene expression and protein localization. BioTechniques, 19(4), 650–655.
11. Kimura, A., & Kobayashi, E. Imaging studies using reporter-gene transgenic rats. Encyclopedia of neuroscience (pp. 97-102)
12. Tung, J. K., Berglund, K., Gutekunst, C., Hochgeschwender, U., & Gross, R. E. (2016). Bioluminescence imaging in live cells and animals. Neurophotonics (Print), 3(2), 025001. doi:10.1117/1.nph.3.2.025001
13. Garcia, H., Lee, H., Boedicker, J., & Phillips, R. (2011). Comparison and Calibration of Different Reporters for Quantitative Analysis of Gene Expression. Biophysical Journal, 101(3), 535-544. doi: 10.1016/j.bpj.2011.06.026
14. Schafhauser, B., Kristofco, L., de Oliveira, C., & Brooks, B. (2018). Global review and analysis of erythromycin in the environment: Occurrence, bioaccumulation and antibiotic resistance hazards. Environmental Pollution, 238, 440-451. doi: 10.1016/j.envpol.2018.03.052
15. Asal, M., Özen, Ö, Şahinler, M., Baysal, H. T., & Polatoğlu, I. (2019). An overview of biomolecules, immobilization methods and support materials of biosensors. Sensor Review, 39(3), 377-386. doi:10.1108/SR-04-2018-0084
16. Bhardwaj, T. (2014). A Review on Immobilization Techniques of Biosensors. International journal of engineering research and technology, 3.
17. Zhao, C., Ge, Z., & Yang, C. (2017). Microfluidic Techniques for Analytes Concentration. Micromachines, 8(1), 28. https://doi.org/10.3390/mi8010028
18. Gaspar, C., Sikanen, T., Franssila, S., & Jokinen, V. (2016). Inkjet-printed silver electrodes on macroporous paper for a paper-based isoelectric focusing device. Biomicrofluidics, 10(6), 064120. https://doi.org/10.1063/1.4973246
19. Krishnamoorthy, G., Wolloscheck, D., Weeks, J., Croft, C., Rybenkov, V., & Zgurskaya, H. (2016). Breaking the permeability barrier ofEscherichia coliby controlled hyperporination of the outer membrane. Antimicrobial Agents And Chemotherapy, AAC.01882-16. doi: 10.1128/aac.01882-16