Overview

We propose two strategies for trace chemical detection and biomarkers in a broad spectrum. We decide to utilize a protein-binding way of detection, leading us to better harness the advantages of shared motifs amongst proteins binding with small molecules. This binding manner ensures the coverage of the target's analogs which may possess similar features, and increase the modularity even more.

Chimeric receptors are assembled for optimal compatibility, and the orthogonal quorum sensing module is responsible for luminescence, optimized by modeling. Moreover, we carefully designed the safety module for a semi-container, minimizing the possible leakage or contamination. Lastly, we perform computational strategies to improve the modules further and put forward new systems we intend to implement in future practice.

The overall genetic circuit

Original Threshold Guard Switch

Schematic diagram of threshold guard switch (XylS system)

We placed the digitalized module downstream of our detection fragment. Here are the critical components of this post-transcriptional control circuit:

The central element is a cis-repressing small RNA (sRNA) based on a naturally occurring transcript found in E. coli. This regulatory sRNA consists of a scaffold sequence and a target binding sequence. The scaffold sequence is derived from the MicC consensus secondary structure, known to recruit the RNA chaperone Hfq. The Hfq chaperone facilitates the interaction between sRNA and target mRNA, leading to mRNA degradation (Yoo et al., 2013).

The second part of the sRNA replaces the natural MicC target binding region and is specifically designed to inhibit the translation of LacI, a transcriptional repressor protein. This 24-base fragment corresponds to the antisense sequence of LacI's translation initiation region, starting from its initiation codon. A LacI-dependent PA1/O4s promoter regulates the production of this sRNA. We chose this specific version of the Plac promoter because it has the best combination of interaction between RNA polymerase (RNAP) and the promoter, along with the operator's position, resulting in the highest level of repression by LacI (Lanzer & Bujard, 1988).

A translational coupler cassette is necessary to control the GOI in cooperation with the upstream LacI repressor gene, which is the target of the sRNA. The capacity of translating ribosomes to unfold mRNA secondary structures serves as an effective technique for coupling translation. By creating a supplementary mRNA structure with a His-tag sequence appended to the 3' end of the LacI gene, we can accomplish translational coupling by blocking the RBS of msf•GFP in our design. In order to block ribosomal recruitment and consequently translation, the sequence was created to form a sturdy hairpin with the Shine-Dalgarno sequence upstream of the GOI (Mendez-Perez et al., 2012).

By replacing the repressor-promoter with different repressor strenth pairs (such as PhIF-PhIFO, LacI-LacIO, TetR-TetRO and AraC-AraCO), we are able to adjust the ultimate threshold value, therefore better suitable for distinct detection needs. The proof of concept of this mechanism and experimental verification is further discussed in our Model and Parts page.

BPA Detection Module

BPA Transformation

BPA can bind to estrogen receptors in the human body, however, since E. coli has no estrogen or similar receptors. We must degrade BPA into other substances for detection.

BPA is known for its resistance to most microorganisms due to the isopropylidene bridge connecting two phenol rings. However, a strain of Pseudomonas putida, known as YC-AE1, has all the necessary enzymes to break it down. In the YC-AE1 genome, three crucial genes encode for three essential enzymes involved in the process (Eltoukhy et al., 2022).

Schematic diagram of BPA conversion to 4-HBA

The first step involves two genes called bisdA and bisdB, which encode ferredoxin and cytochrome P450, respectively. These enzymes efficiently convert BPA into either 1,2-bis(4-hydroxyphenyl)-2-propanol or 2,2 bis(4-hydroxyphenyl)-1-propanol.

After this, the intermediates are transformed into p-Hydroxybenzaldehyde using common dehydratase enzymes found in prokaryotes.

The third essential gene, hbd, encodes a dehydrogenase that converts p-Hydroxybenzaldehyde into p-Hydroxybenzoic acid (4-HBA in this text), which is the substance that we detect. This pathway is unique from other pathways, ensuring our detection methods' accuracy.

Previous studies have confirmed that this pathway is functional in E. coli, transforming BPA to 4-HBA (Jia et al., 2020).

PobR Threshold Guard Switch

The gene pobR creates a transcriptional activator that attaches to the pobR operator on the dsDNA before combining with 4-HBA. Once 4-HBA is introduced to the solution, PobR binds with it and triggers the transcription of the dual pobA/R promoter on the side of pobA.

One of the advantages of this protein is its sensitivity and low leakage properties. Even tiny amounts of 4-HBA, at the micromolar level, can trigger transcription. This characteristic is crucial in creating a high-quality digitizer with a sharp response between two stable states(Calles et al., 2019).

Additionally, research indicates that most analogs of 4-HBA, such as p-aminobenzoate, can impede the activation of PobR, ensuring precise detection of 4-HBA.

Moreover, the combination of the non-activated PobR and pobR operator will inhibit the transcription of pobR when there is no 4-HBA stimulus, reducing the pressure on our bacteria.

Given PobR's excellent properties, we create a chimeric switch by combining it with a portion of the original threshold guard switch.

Schematic diagram illustrating how we create the PobR threshold guard switch

The original pobA/R operon comprises six main components: pobA, pobA RBS, pobA/R dual-directional promoter, pobR operator, pobR RBS, and pobR. The pobA gene encodes an unnecessary enzyme that breaks down 4-HBA. Therefore, we remove the pobA and pobA RBS components but retain the entire promoter, pobR RBS, and pobR.

Regarding the original threshold guard switch, we eliminate the XylS system, which includes the XylS protein and its corresponding promoter. We then combine these two sections and anticipate that the new chimeric digitizer will function properly.

We use GFP to test our new chimeric digitizer. We will replace GFP with luxI to establish our bio-sensor if it functions well.

Bile Salts Detection Module

Vibrio cholerae is an enteropathogenic bacteria that uses bile salts as an intestinal location signal and, therefore, presents as a native capable biosensor for bile salts. In Vibrio cholerae, TcpP-TcpH is mainly responsible for primary bile acid detection. By rewiring the TcpP-TcpH part into our engineered bacteria, we could achieve bile salt detection with high sensitivity and a short waiting time.

TcpH is a cofactor necessary for TcpP to function, and TcpP-LBD is a single-domain protein that binds the human primary bile acid and could form a dimer when bound. This is achieved by rewiring TcpP–TcpH from V. cholerae to E.coli and fusing it with the CadC DNA binding domain to construct a chimeric receptor. When activated, TcpP-CadC forms a dimer and activates the pCadBA promoter, synthesizing LuxI and turning on the quorum sensing cascade.

Since TcpP-LBD is a single-domain protein, it could be easily swapped into other ligand binding domains like nanobodies, promoting compatibility for the platform.

Schematic sketch of the mechanism of CadC-TcpP

We transplanted the TcpP-CadC module into the digitalizer, thus achieving threshold control and anti-leakage simultaneously, as is shown below.

Demonstration of rewiring CadC-TcpP into threshold guard switch

Orthogonal Quorum Sensing

Schematic sketch of the mechanism of quorum sensing

Quorum sensing is a mechanism employed by a diverse range of bacteria to modulate the expression of genes in a manner that is contingent upon the density of cells. Bacterial populations can coordinate their behavior by utilizing tiny molecules known as autoinducers. These autoinducers are synthesized by cognate synthases and afterward detected by particular receptors. N-acylated homoserine lactones (AHLs) are a class of autoinducers. AHL-based quorum sensing requires an AHL-synthase and a LuxR-type regulator whose activity is modified by the AHL. After AHL binds to LuxR, LuxR dimerizes and binds to DNA, resulting in downstream transcription of pLuxm.

Our initial plan was to create a genetic circuit that featured a luxI downstream of pLuxm for positive feedback. However, recent research has suggested that the quorum sensing system would perform better without this downstream luxI (Hui et al., 2022). Therefore, we decided to verify the conclusion's correctness and then chose the higher efficiency pathway. The autoinducers, AHL, produced by the upstream system, can diffuse through the cell membrane into the growth medium. Here, we used a mutated LuxR–lux box system called LuxRm-Pluxm. LuxR synthesized by E. coli can not recognize the pLuxm promotor. Compared with the wild-type LuxR–lux box system, the transcriptional activities of LuxRm-pLuxm are better, and the background expression is more tightly regulated (Lu, 2016).

The J23119 promoter drives the expression of the luxRm gene continuously. If the upstream pathway is closed, AHL concentration is low, only showing slight leakage. As the upstream pathway opens, the medium's autoinducers start accumulating in a confined environment. When enough AHLs have accumulated in the medium, they can enter the cell, directly binding the LuxRm protein to activate pLuxm and downstream transcription.

Nanoluc

We utilize Nanoluc as a characterization of our final output. Nanoluc reacts with exogenous furimazine and emits bioluminescence, eliminating the background noise and giving us a more comprehensive range of linear spectrum correlations. Moreover, this bioluminescence is strong enough after the amplification and could be more easily detected. The signal emitted by Nanoluc also follows a predictable pattern(as demonstrated in Results), which is more convenient to characterize and analyze.

Since we are detecting it in semi-contained Hardware, adding furimazine in the reaction chamber should be feasible. The photoresistor will directly detect the luminescence, thus converted to electric signals, and readily available for further analysis(as demonstrated in Hardware).

Demonstration of the mechanism of nanoluc

Kill Switch

Doc-Phd Toxin-antitoxin System

For the final version of the kill switch, we decided to use an induced circuit. We had initially considered using an unnatural amino-acid-based kill switch, but we had to abandon that idea due to financial and time constraints. Additionally, we plan to design a temperature-controlled kill switch, detailed in the temperature-sensitive kill switch section.

The inducer of our kill switch is IPTG. With IPTG present, the bacteria can survive, but if it leaks into the environment without it, the bacteria will die.

We chose the Doc toxin(BBa_K3392000) and its antidote Phd(BBa_K3143002). As a result, the genetic circuit can be further simplified more straightforwardly and conveniently. The J23119 promoter was utilized to facilitate the expression of the Doc gene, while the Plac promoter was employed to regulate the expression of the Phd gene. The outcomes obtained were consistent, whereby the presence of IPTG in the system led to the expression of the Phd gene, resulting in its combination with the toxin. Conversely, in the absence of IPTG, the bacteria were released, activating the Doc gene and subsequent bacterial cell death.

Genetic circuit of the kill switch

Unnatural Amino Acid System

Since we mostly contain our engineered bacteria in the hardware, utilizing substance constraints is feasible. Therefore, we use a toxin-antidote system relying on an unnatural amino acid, 3-iodo-L-tyrosine(IY). Only when fed with IY can the engineered bacteria stay alive, which can be added to the hardware. If the bacteria unexpectedly escape or leak, it cannot survive without IY, leading to automatic death.

This is achieved using an amber stop codon and an amber-specific acyl-tRNA synthase. Colicin E3(colE3) is a highly toxic RNase that kills the host bacterium with a few molecules. 1amb-IMME3 could directly inhibit the activity of Colincin E3 by forming a complex with it. When provided with IY and the expression of Full IY-tRNA synthase and amber suppressor tRNA, bacteria could incorporate 3-iodo-L-tyrosine(IY) into the antidote. Therefore, the engineered bacteria can stay alive only with foreign supplements of IY.

Demonstration of the mechanism of unnatural amino acid system

Rational Design

In the orthogonal quorum sensing part, we tested the mutated LuxR's capability of quorum sensing and orthogonality(see page Results). Moreover, we perform computational methods to gain more insights into the structural basis of the LuxR-VAI interaction.

We first searched for the structure of LuxR and VAI, performed molecular docking, and analyzed the interaction between VAI and the residues of amino acids(see page Results).

VAI-LuxR interaction predicted by Molecular Docking

Further Improvement

Self-renewal System

We want the E. coli to obtain the ability to express LuxI at a basic level to enable the start of quorum sensing. However, when the system is working, the basic level expression of LuxI ought to be terminated, and during the restore period, LuxI should be expressed again.

As the working temperature is 37℃, and the restoring temperature is 25℃, we devised the idea to use a temperature-sensitive RBS to control the expression of LuxI. This RBS was originally from iGEM20_UNILausanne, named heat-repressible RNA-thermosensor F2.

When the temperature is high(37℃), the RNase E cleavage site(RC) of this part will be cleavaged by RNase E, turning off the downstream expression. On the other hand, at a low temperature(27℃), the Anti-RNase E cleavage site(ARC) can complement with the RC, and thus, the RNase cannot bind and cut the strain, ensuring the complete element to allow the downstream gene to be expressed.

As a result, the LuxI can be expressed at the restoring temperature and turned off at the working temperature.

Population Density Control

The density of the bacteria can influence the fluorescence intensity, so uncontrolled thickness can interfere with the detection signal. A circuit to control the bacteria density is considered for semiquantitative detection.

The phosphotransferase system (PTS) is widely found in fungi and archaea. It mainly phosphorylates various sugars and their phosphorylation through phosphoric acid cascade reactions. It plays a vital role in sugar consumption and sugar metabolism in organisms. The acidic phosphate carrier protein HPr and mutant LsrK can form a co-crystal structure, inhibiting the activity of LsrK and unable to phosphorylate AI-2 normally, preventing the regular progress of quorum sensing. When bacteria are in a stable growth state, HPr is phosphorylated, and its binding ability to LsrK is reduced. It can phosphorylate AI-2 normally, making quorum sensing normal. Based on the close connection between quorum sensing and sugar sensing mentioned above, Stephens et al. verified that induction of HPr expression can increase cell growth rate and used orthogonal AI-1 and AI-2 quorum sensing systems to construct an engineering consortium and designed a system that responds to AI-2 signals and generates translation products of AI-1 signal molecules and the control system that regulates the growth rate of bacterial colonies based on AI-1 signal levels have developed a method for autonomously controlling the sparse density of engineered bacteria.

Demonstration of the mechanism of population density control

Temperature-Mediated Switch

In addition to the two kill switches mentioned above, we have proposed a prospective strategy to develop a temperature-regulated kill switch. In this proposed design, we aim to create a component responsive to temperature variations, explicitly functioning at elevated temperatures and reverting to its original state at lower temperatures within the facility. The designated operational temperature is set at 37℃, while the restoration temperature is set at 25℃. When the temperature falls outside this specified range, the kill switch will be activated, leading to the termination of the engineered bacteria. For more information, please see the page Safety.

End-to-End Characterization

Many quantitive biological sample tests, such as BCA, require a standard curve. This is due to the variation of reference reaction conditions. Although we possess a multi-sampling ability to draw a standard curve, we still firmly believe that direct quantification will be highly thrilling.

To achieve this in a relatively complex system, we must strictly control the reaction chamber(temperature control) and the active reporting biosensors(density control). With the assistance of ambitious modeling, we aim to achieve an end-to-end detection result.

In other words, we could use an improved strategy to read luminescence and directly convert it into the absolute concentration of the sample. We will consistently work on it and strive to gain even more robustness in creating our whole biosensor platform.

Reference

1. Calles, B., Goñi-Moreno, Á., & De Lorenzo, V. (2019). Digitalizing heterologous gene expression in Gram‐negative bacteria with a portable ON/OFF module. Molecular Systems Biology, 15(12). https://doi.org/10.15252/msb.20188777
2. Na D, Yoo SM, Chung H, Park H, Park JH, Lee SY (2013) Metabolic engineering of Escherichia coli using synthetic small regulatory RNAs. Nat Biotech 31: 170 – 174
3. Lanzer M, Bujard H (1988) Promoters largely determine the efficiency of repressor action. Proc Natl Acad Sci USA 85: 8973 – 8977
4. Mendez-Perez D, Gunasekaran S, Orler VJ, Pfleger BF (2012) A translationcoupling DNA cassette for monitoring protein translation in Escherichia coli. Metab Eng 14: 298 – 305
5. Eltoukhy, A., Jia, Y., Lamraoui, I., Abo-Kadoum, M. A., Atta, O. M., Nahurira, R., . . . Yan, Y. (2022). Transcriptome analysis and cytochrome P450 monooxygenase reveal the molecular mechanism of Bisphenol A degradation by Pseudomonas putida strain YC-AE1. BMC Microbiol, 22(1), 294. doi:10.1186/s12866-022-02689-6
6. Jia, Y., Eltoukhy, A., Wang, J., Li, X., Hlaing, T. S., Aung, M. M., . . . Yan, Y. (2020). Biodegradation of Bisphenol A by Sphingobium sp. YC-JY1 and the Essential Role of Cytochrome P450 Monooxygenase. Int J Mol Sci, 21(10). doi:10.3390/ijms21103588
7. Li, Z., & Nair, S. K. (2012). Quorum sensing: How bacteria can coordinate activity and synchronize their response to external signals? Protein Science, 21(10), 1403-1417. https://doi.org/10.1002/pro.2132
8. Lu, Y. (2016). Engineering Vibrio fischeri transcriptional activator LuxR for diverse transcriptional activities. Biotechnology Letters, 38(9), 1459-1463. https://doi.org/10.1007/s10529-016-2134-z
9. Hui, C. Y., Hu, S. Y., Li, L. M., Yun, J. P., Zhang, Y. F., Yi, J., . . . Guo, Y. (2022). Metabolic engineering of the carotenoid biosynthetic pathway toward a specific and sensitive inorganic mercury biosensor. RSC Adv, 12(55), 36142-36148. doi:10.1039/d2ra06764a
10. Vaaben, T. H., Vazquez-Uribe, R., & Sommer, M. O. (2022). Characterization of eight bacterial biosensors for Microbial Diagnostic and Therapeutic Applications. ACS Synthetic Biology, 11(12), 4184–4192. https://doi.org/10.1021/acssynbio.2c00491
11. Chang, HJ., Zúñiga, A., Conejero, I. et al. Programmable receptors enable bacterial biosensors to detect pathological biomarkers in clinical samples. Nat Commun 12, 5216 (2021). https://doi.org/10.1038/s41467-021-25538-y