Engineering

1. Design

Identifying the local problem and solution - our inspiration

After interviewing various experts and Hong Kong government bodies, we discovered that a major source of heavy metal contamination is the improper disposal of building and construction materials by workers. This leads to heavy metals leaking into Hong Kong's water sources, directly affecting the local population. Furthermore, we noticed that traditional Chinese medicine is widely used in Hong Kong, and there have been reports of heavy metal contamination, especially zinc and mercury, in some herbal medicines. This discovery highlighted the importance of developing a convenient and affordable method to detect these pollutants.

Taking these factors into account, we decided to focus our project on developing biosensors for detecting heavy metals using synthetic biology. Biosensors offer several advantages, including being more cost-effective, environmentally friendly since they do not require toxic chemicals for sample preparation, and user-friendly due to the lack of special equipment or advanced techniques needed for operation. Additionally, their portability allows for convenient on-site detection. After conducting research, we identified lead, cadmium, mercury, and zinc as the most prevalent and hazardous heavy metal pollutants in Hong Kong.

As a result, our project aims to create biosensors specifically designed to detect these four heavy metals.

Designing a workflow that can facilitate problem-solving

Our primary goal is to develop a biosensor that can be visually observed with the naked eye, allowing individuals without specialized equipment or technical expertise to easily interpret the results—even in settings like our school laboratory, which lacks advanced instruments like a spectrometer.

Although our initial research on biosensors revealed that fluorescent proteins like GFP and RFP, as well as luminescent signals from luciferase, are often more sensitive and responsive, we chose to use chromoproteins. These proteins produce visible colour changes in E. coli, making them ideal for our biosensor design.

To begin, we conducted a preliminary study of various chromoproteins available in the iGEM registry. Thanks to the hard work of previous iGEM teams, the registry offers a long and detailed list of chromoprotein candidates, along with valuable information about each. However, we weren"t sure if the expression and culturing conditions in our basic laboratory setup would produce the same results as those reported in the registry. Therefore, we decided to test and briefly characterize the chromoproteins ourselves as the first part of our project.

From the range of colours available, we selected chromoproteins that offered the greatest contrast with the natural colour of bacterial cultures. Green, yellow, and orange proteins were too similar to the pale or yellowish colour of the broth we use, so we eliminated those from consideration. Instead, we chose amilCP (BBa_K592009) (dark blue), cjBlue (BBa_K592011) (dark green), tsPurple (BBa_K1033906) (bright purple), dTomato (BBa_K4813000) (sharp pink), and eforRed (BBa_K592012) (dark red) due to their high contrast and bright colours, as reported in the literature.

Next, we designed five expression constructs for these chromoproteins, utilizing constitutive promoters (BBa_J23100) and a strong ribosome binding site (BBa_B0032). We recognize that using strong promoters might not fully represent the expression conditions of our final metal biosensor constructs, but the results will provide a rough reference for further investigation.

More importantly, we recognize that problem-solving is the most challenging and crucial aspect of the engineering cycle. Therefore, we created these designs to serve several key purposes as well:

1. Ensuring our bacterial strain, lab conditions, and experimental setup can successfully express these proteins.
2. Verifying that the colour of the expressed proteins matches the expected results from previous reports.
3. Using these five constructs to verify that our experiment design, protocols and materials are correct and functional.
4. Employing these constructs as positive controls in future experiments, providing a reference point for troubleshooting any issues that arise when we create the biosensor constructs.

In our later experiments, these constructs proved to be invaluable for troubleshooting and ensuring that our cloning process was smooth and effective. They provided us with a reliable reference point, helping us confirm that our cloning techniques were working as intended.

2. Build

Cloning and characterization of chromoprotein expressions

First, we designed constitutive expression constructs for the chromoproteins amilCP(BBa_K5152002), cjBlue (BBa_K5152001), tsPurple (BBa_K5152000), dTomato (BBa_K4813005), and eforRed(BBa_K5152003). For the dTomato part, it was created by our previous iGEM team, we further validated its functions and expressions in this project and documented it in the part registry (BBa_K4813000).

The graphic representation of these constructs is shown below.

The chromoprotein CDS (coding sequences) are driven by the strong Anderson family constitutive promoter (BBa_J23100), paired with a strong RBS (BBa_B0034), and terminated by a strong double terminator (BBa_B0015). Additionally, the constructs include a 5' 20 nt overlapping region and a 3' 20 nt overlapping region to facilitate the HiFi assembly process with the pUC19 vector, using the restriction sites PstI and EcoRI. We opted for double enzyme digestion to linearize the plasmids, as this reduces the likelihood of self-ligation during the assembly process.

The pUC19 vector was sourced from the glycerol stocks of our previous project. It was purified using a miniprep procedure and then subjected to restriction digestion for 2 hours. After digestion, the product was run on a gel electrophoresis, and the band corresponding to the vector (as illustrated in the diagram below) was extracted and purified using a gel purification kit.

Figure 1: Gel photo showing the linearized pUC19 vector (~2.6 kbp) after digestion with PstI and EcoRI. The highlighted band corresponds to the digested vector, which will be extracted and purified using a gel purification kit.

Since we do not have access to spectrophotometry machines due to their high cost, we estimated the concentration of DNA in our samples by comparing the brightness intensity of the bands with the DNA ladder on the gel. Based on this estimation, we calculated the ratio of insert to vector to be approximately 3:1 and proceeded with the assembly process. The chromoprotein inserts were synthesized as gBlocks from IDT.

The five chromoprotein expression constructs were then assembled and transformed into TOP10 E. coli strains using the heat shock transformation method. The transformants were selected on LB agar plates containing ampicillin.

After 16 hours of incubation, most of the colonies showed significant colouration corresponding to their constructs, except for cjBlue. However, after incubating for more than 24 hours, cjBlue also began to express its colour.

Figure 2: E. coli colonies transformed with the constitutive chromoprotein expression constructs (cjBlue, tsPurple, amilCP, eforRed, and dTomato) on LB/AMP selection plates.

Figure 3: E. coli cells transformed with the cjBlue expressing construct show a strange pattern of chromoprotein expression after 16 hours of incubation at 37°C.

Figure 4: Cell pellets of E. coli expressing different chromoproteins obtained by centrifugation (8000 g, 2 minutes) of 1 mL of cell culture. Each pellet displays the characteristic colour of the expressed chromoprotein.

We inoculated the colonies from all the plates and repeated the spreading process. The same results were observed, confirming that cjBlue expression was not working as expected in our laboratory conditions. As a result, we decided to exclude cjBlue and selected the remaining chromoproteins as candidates for our biosensor construction. Additionally, since dTomato has been successfully utilized in previous projects by our lab, we are familiar with its expression properties, and it will be used as a positive control in future cloning processes.

3. Test

Using the workflow and conditions as a positive control

With the successful expression of the chromoproteins, we proceeded to further characterize their expression properties, specifically focusing on the time required for visible result reading.

We cultured the cells and harvested them at different time points: 4 hours, 8 hours, 12 hours, 18 hours, and 24 hours. At each time point, we collected 1 mL of the culture and centrifuged it at 8,000 g for 2 minutes to obtain the cell pellets for observation.

The results at 12 hours and 18 hours time point are shown in the diagrams for reference below.

Figure 5: The cell pellets after 12 hours and 18 hours of incubation at 37°C. The pellets after 18 hours of incubation exhibit significantly higher colour intensity compared to those at 12 hours. pUC19 represents the cells transformed with the pUC19 empty vector as a control.

We recognize that using a constitutive promoter might not perfectly reflect chromoprotein expression when transitioning to metal-inducing promoters. However, this data provides a useful preliminary range for our study.

Additionally, we cultured the cells at room temperature without shaking, which resulted in a slower colour change. Noticeable colour appeared only after at least 24 hours of incubation.

4. Learn

From our results, we drew several key conclusions:

1. Influence of Culture Conditions: The expression of chromoproteins is highly affected by culture conditions and may not always match the results reported in the literature. In our experiments,cjBlue consistently only showed expression after 24 hours of incubation, making it impractical to use as a reporter gene in our project. As a result, we decided to exclude cjBluefrom our biosensor design.
2. Optimizing Expression with Temperature and Shaking: We observed that incubating the cultures at 37°C with shaking significantly enhanced chromoprotein expression. This discovery led us to collaborate with our hardware team, requesting them to integrate temperature control and shaking functions into our hardware device to optimize the performance of our biosensor.
3. Time Required for Result Reading: We found that a minimum of 12 hours is required for the colour intensity of the cells to provide a clear contrast for result reading. While longer incubation times (around 18 hours) produce more obvious results, we also discovered that in cells with metal-sensing constructs, extended incubation can lead to leaky expression of the chromoproteins, resulting in false positives. Therefore, it is essential to establish a specific time frame and a threshold for measurement in our hardware device to ensure accurate detection without prolonged incubation.

After confirming that the chromoproteins have stable expression and that our measurement method is reliable, we moved on to the next phase of our project: designing our biosensor devices. This step involves incorporating the validated chromoprotein reporters into constructs that respond to the presence of the target heavy metals, ensuring accurate and practical detection.

5. Design

Designing Heavy Metal-Inducing Expression Constructs

With the knowledge gained, we proceeded to create our heavy metal biosensors.

To achieve this, we conducted a thorough literature review on metal-inducible promoters and evaluated their potential for use in our project. By studying various sources and exploring the BioBrick registry, we identified several promising candidates for our biosensor design, including:

1. pPbr operon (BBa_I721001) (Lead-inducible)
2. pCadA operon (BBa_K1724000) (BBa_K174015) (Cadmium-inducible)
3. pYodA operon (BBa_K896008) (Cadmium-inducible)
4. pMerT operon (BBa_K346002) (Mercury-inducible)
5. pZnt operon (BBa_K346002) (Zinc-inducible)

Characterizing and Designing Biosensor Constructs

We then analyzed the characteristics of each operon to tailor our biosensor constructs to meet the goals of our project.

For pYodA, pZnt, and pCadA operons, the literature reported that they are native operons in E. coli, meaning there is no need to ectopically express their regulatory proteins.

However, for the pPbr, a synthetically modified pCadA, and pMerT operons, which are not naturally found in E. coli, we needed to co-express their regulatory proteins:

• PbrR repressor (BBa_I721002) for pPbr (lead-inducible)
• A modified cadmium-sensitive MerR repressor (BBa_K174015) for the pCadA operon
• Mercury-binding MerR (BBa_K346002) activator for pMerT

The illustrations of the regulatory mechanisms of the above operons are shown below:

1. pPbr operon (Positive regulation)

2. MerR-pCadA operon (Negative regulation)

3. MerR-pMerT operon (Positive regulation)

Balancing Expression of Regulatory Proteins

As mentioned, when expressing regulatory proteins, we carefully considered expression strength. Overexpressing a repressor can reduce the biosensor's sensitivity, while under-expressing it may cause leaky expression and false positives. Conversely, overexpressing an activator can lead to false positives, while under-expressing it results in low sensitivity.

After designing the constructs, we initially had 7 designs ready. We ordered these constructs in the form of gBlocks from IDT. However, we encountered an issue with one of the lead sensor constructs—due to the presence of many repeating sequences, IDT was unable to synthesize it.

The 6 final designs of our constructs are illustrated below:

1. PbrR-pPbr lead sensing with dTomato reporter construct

2. MerR-pCadA cadmium sensing with amilCP reporter construct

3. pCadA cadmium sensing with amilCP reporter construct

4. pYodA cadmium sensing with amilCP reporter construct

5. MerR-pMerT mercury sensing with tsPurple reporter construct

6. pZnt zinc sensing with eforRed reporter construct

At this stage, everything seemed to be running smoothly, but soon we faced several major challenges that required significant time and effort to resolve. In the following section, we will discuss the hurdles we encountered and the problem-solving process that helped us overcome them.

Experiment plans

1. Correlation Between Chromoprotein colour Intensity and Heavy Metal Concentration: We planned to quantify the relationship between the intensity of the chromoprotein colour and the concentration of the corresponding heavy metal, allowing us to establish a calibration curve for more precise measurements.
2. Specificity of the Biosensors: We intended to test the specificity of our biosensors to ensure they only respond to their target heavy metals and not to other metals, preventing false positive results.
3. Interaction of Mixed Biosensors: We wanted to investigate what happens when different biosensors are mixed together. Specifically, we aimed to see if they can function simultaneously and detect different metals in the same environment without interfering with each other.
4. Demonstration in our hardware device: Finally, we planned to demonstrate the functionality of our biosensors in the auto-reporting machine developed by our hardware team, to show how the sensors could be incorporated into a fully automated monitoring system.

However, due to time constraints, we were only able to complete part of these experiments. Most of the planned investigations will be carried out in our future studies.

6. Build

Cloning attempts of metal-inducing expression constructs

Upon receiving the gBlocks from IDT, we immediately began the cloning process. Building on our experience from cloning the constitutive chromoprotein expression strains, we followed the same protocol we had successfully used before.

During our first cloning attempt, we encountered an issue with lawn growth on all plates. (Fig. 6) After investigation, we discovered the problem was due to the ampicillin in our plates. We found that our ampicillin might have expired, as it was stored at -20°C instead of -80°C and had been stocked for over a year.

Figure 6: The plates showing lawns formed after the first cloning attempt. Upon investigation, we discovered that the issue was caused by the absence of ampicillin in the plates, which allowed non-selective bacterial growth.

To address this, we re-spread the remaining transformants onto a new set of plates containing the appropriate ampicillin concentration. This ensured that only the transformed colonies with the correct plasmids would survive and grow.

Figure 7: The plates re-spread with the remaining transformants. No colonies were formed, prompting us to repeat the assembly process and eventually verify the DNA sequence to identify potential issues with the cloning.

After incubation, all plates were empty with no colonies, except for the positive controls. We initially suspected that the transformants had been stored in the fridge, which may have reduced their viability, preventing colony formation.

In response, we repeated the assembly process and transformed the products once more. This time, a few colonies appeared on the plates. After consulting with our teacher, we learned these colonies were most likely false positives. However, at that time, we believed that these colonies were successful, so we proceeded with colony PCR to verify the clones.

Unexpectedly by us, all the colonies from the colony PCR showed no bands, indicating that there was no insert in the plasmids. Additionally, we made a significant mistake during the PCR process by not adding positive controls to the reaction, which could have helped us interpret the results more accurately.

Figure 8: One of the gel results from the colony PCR of the first cloning attempt, showing no bands formed. This indicates that none of the colonies contained the desired insert we attempted to assemble. The PCR was run using the M13 standard primer for pUC19 vectors. The faint bands at the bottom of the gel represent the primers.

Learned from this experience, we had consistently used a positive control setup during the assembly process. These positive control assemblies were always successful to show colonies on the plates, which led us to suspect that the issue might be with the PCR process itself. As a result, we conducted an additional round of PCR verification to troubleshoot the problem further.

Validation of PCR

From the PCR results, we confirmed that our PCR process was functioning correctly. After consulting with our teacher, we also understood the critical importance of including positive controls in our PCR setups. This would help us better interpret the results and avoid unnecessary confusion in future experiments.

Figure 9: Gel photo showing the result of validating that our PCR conditions and machine are functioning properly. The DNA template used was from our team's previous iGEM project, and the primers were the same M13 primers used in our current project, under the same PCR conditions. All the band sizes are within the expected range, confirming that the PCR setup is working correctly.

Next, we needed to investigate other potential causes for the cloning failure. Since the PCR process was confirmed to be working properly, we shifted our focus to other aspects of the workflow to identify any issues.

Identified the mistakes

Since both the cloning process and PCR were verified to be error-free, we returned to the initial step and double-checked the DNA sequence we ordered. After careful review, we discovered that the 5' overlapping sequence was incorrectly designed to overlap with the EcoRI digestion site instead of PstI, causing the assembly process to fail.

As a result, we had to immediately re-design the constructs with the correct overlapping sequences and reorder the DNA. This step was crucial to ensuring the success of the cloning and moving forward with our project.

The successful cloning attempt of metal-inducing expression constructs

After receiving the newly synthesized DNAs from IDT and TwistBioscience (many thanks to them for their kind sponsorship!), we immediately proceeded with the assembly process again. Using the same workflow that we had successfully employed for the constitutive constructs, this time we managed to obtain colonies from five out of the six assembled constructs. These included:

1. Lead-inducing pPbr with the co-expression of PbrR activator construct (BBa_K5152004)
2. Cadmium-inducing pCadA with the co-expression of modified MerR repressor construct (BBa_K5152006)
3. Cadmium-inducing pYodA construct (BBa_K5152007)
4. Zinc-inducing pZnt construct (BBa_K5152009)
5. Mercury-inducing pMerT with co-expression of MerR activator construct (BBa_K5152008)

Following this, we proceeded with colony PCR to verify the presence of the correct inserts in the clones.

Figure 10: We tracked the clones by picking and inoculating them on new agar plates. As shown in the picture, some of the clones exhibited leaky expression (visible colour without induction). To avoid potential issues with false positives, we decided not to select clones showing colour for further study.

Clones verified by colony PCR

From the gel photos below, some of the colony PCR results showed bands corresponding to the expected sizes of the inserts. Therefore, we selected these colonies for our subsequent functional studies, ensuring that only the correctly assembled constructs would be used for further experiments.

Figure 11:

(a): PbrR construct colony PCR result. The expected band size of ~1.5 kbp was observed in lanes 3, 4, and 5.

(b): MerR-pCadA construct colony PCR result. The expected band size of ~1.6 kbp was observed in lanes 4, 5, and 6.

(c): pZnt construct and pYodA colony PCR result. The expected band size of ~600 bp was observed in the corresponding lanes.

Despite some successful colony PCR results, we encountered several unexpected results that raised concerns. Specifically, some colonies showed no insert, which was unexpected given the steps we took to minimize self-ligation:

Gel purification was done after double enzyme digestion of the pUC19 vector, significantly reducing the chance of self-ligation. Thus, there should not have been any clones carrying antibiotic resistance unless they contained the assembled plasmids.

However, several colonies lacked inserts, which led us to consider potential explanations:

1. Plasmid contamination during the transformation process, despite our efforts to prevent contamination.
2. Inappropriate antibiotic concentration in the selection plates. To investigate this possibility, we tested our ampicillin plates by spreading them with E. coli carrying the pUC19 vector and E. coli without resistance.

Nevertheless, we proceeded by selecting the colonies that showed correct band sizes in the colony PCR for our functional assays.

Functional studies of the E. coli heavy metal biosensors

One key observation from our transformed cells was that some exhibited leaky expression of the chromoprotein even without metal induction, especially when incubated for too long (typically ≥ 18 hours). To address this, we carefully selected colonies that did not show leaky expression for the functional assays and kept the measurement time within 18 hours in all subsequent experiments.

After overcoming several hurdles, we finally obtained the correct clones. However, due to previous errors in DNA design, equipment limitations, and various challenges during the experiment, the process was delayed. For example:

• We lacked a refrigerated centrifuge, making it impossible to prepare our own competent cells. Thankfully, the Amgen Biotechnology Experience Program Hong Kong generously sponsored us with competent cells when we ran out.
• It was already September by the time we had the correct clones, leaving us with limited time before the wiki freeze.
• Our small shaking incubator could only hold 12 small conical flasks, which required careful planning to optimize time and space usage as the deadline approached.

Given the time constraints, we decided to focus on proving that all our constructs could respond to the corresponding heavy metals. To maximize our chances of success, we conducted research on the concentration of heavy metals used in literature reports and chose the highest reported concentration for each metal that had been shown to induce a response. This approach ensured that the metal concentration would be sufficient to activate the chromoprotein expression and demonstrate proof of concept.

The concentrations of heavy metals used in our experiments are listed in the table below:

We also planned to explore whether the cells could be cultured in 15 mL Falcon tubes, as this would significantly reduce the amount of space required, allowing us to conduct more experiments simultaneously. However, during our trials, we found that this culturing method was not feasible. The specific reasons for this will be discussed in the section below.

After these preliminary tests, we proceeded to use the selected heavy metal concentrations for our experimental setups.

7. Test

Characterization of the biosensors

For each biosensor, our initial experimental setup involved adding the corresponding heavy metal (at the final concentrations listed in the table above) and comparing the results with a control setup where no heavy metal was added.

In the later stages of our investigation, following a suggestion from our teacher, we added an additional control setup: E. coli containing only the pUC19 plasmid, with the corresponding heavy metal added. This control was crucial to demonstrate that the heavy metal alone would not cause any colour change in E. coli without the biosensor constructs.

This additional control was particularly important because we encountered an unexpected issue during our experiments. When we grew the cells in 15 mL Falcon tubes with the heavy metal, the cells exhibited a strange blackened and brown colouration. Concerned that this could cloud our judgment on the experimental results, we conducted a control experiment using E. coli with only the pUC19 plasmid in conical flasks. In these experiments, the unusual colouration was not observed, suggesting that the problem was specific to the Falcon tube culturing method (as shown in the figure below).

Given this observation, we decided to discontinue the use of Falcon tubes for culturing, as the method could lead to misleading results. However, this decision significantly slowed our experimental progress due to the limited space available in the shaking incubator when using conical flasks.

Figure 12:

(a): The biosensor culture grown in a 15 mL Falcon tube with lead (Pb) added. Initially, based on the appearance of the culture, we thought there was a colour change due to the biosensor (BBa_K5152004) successfully detecting the metal. However, this was not the case when we spun down the culture to observe the pellet.

(b): A strange colouration appeared in the cell pellet when incubated in 15 mL Falcon tubes with heavy metal. This unusual colour change could potentially cloud our judgment of the experimental results. Therefore, we decided to discontinue using Falcon tubes for the experimental setup. Notably, this colour change was not observed in E. coli carrying pUC19 when incubated in conical flasks.

We suspect that the unusual colouration observed in the cells grown in 15 mL Falcon tubes with heavy metals is due to poor aeration during incubation. This lack of sufficient oxygen may have created an anaerobic environment, leading to the strange appearance of the cells. However, this is just a hypothesis, and further investigation would be needed to confirm this phenomenon and understand the underlying cause.

After conducting the functional studies, we obtained the following results:

I. Lead Biosensor (BBa_K5152004)

In the lead biosensor, the results were promising. To ensure the reliability of the findings, we repeated the experiment three times.

Upon adding 100 µM of lead and incubating for 12 hours, the experimental setup showed a clear red colouration in the pellet, indicative of dTomato expression, compared to the white pellet in the control. Additionally, a mild red colour was visible in the culture itself, further supporting the successful detection of lead by the biosensor (as shown in the figure below).



Figure 13:

(a): By naked-eye observation, the cultures of the PbrR-pPbr lead biosensor show a distinguishable red colouration compared to the negative control setup, where no lead was added.

(b): The red colouration is even more pronounced in the pellet, highlighting the expression of the reporter gene (dTomato) (BBa_K4813005). This confirms that our biosensor is successfully responding to the presence of lead (Pb).




Although the colour intensity of culture medium after 12 hours of incubation was not significantly different from the control, we predicted that extending the incubation time would result in a more pronounced colour change in the culture medium. This prediction was confirmed (as shown in the figure below).

After 18 hours of incubation, the red colouration in the experimental setup became much more obvious. However, a mild red colour was also observed in the no heavy metal control setup, indicating leaky expression. This leaky expression is a potential issue that we aim to investigate in further experiments.



Figure 14: The PbrR-pPbr lead biosensor shows a significant red colouration after 18 hours of incubation with 100 µM of lead. However, when compared to the pUC19-only control setup, the PbrR-pPbr biosensor without added lead also displayed a mild red colouration, indicating that extended incubation leads to a mild leaky expression of the red chromoprotein.




Next, we investigated whether the expression of the biosensor is concentration-dependent. We added final concentrations of 0.01, 0.1, 1, 10, 500, and 1000 µM of lead nitrate solution to the biosensor cultures.

Without needing to harvest the cells, we observed that the biosensor's response was indeed concentration-dependent within the range of 0.1 to 10 µM. A higher concentration of lead resulted in a more intense colour signal from the reporter. This demonstrates that our biosensor's chromoprotein expression correlates with the lead concentration (as shown in the figure below).



Figure 15: The red colour intensity of the lead biosensor increases with an observable difference as the concentration of lead rises from 0.1 µM to 10 µM. However, when the concentration exceeds 500 µM, there appears to be a negative effect on the colour intensity, potentially due to toxicity or inhibition of cellular functions at high lead concentrations.



Figure 16: When the cultures with different concentrations of lead were spun down into pellets, the colour intensities were all too high, making the differences indistinguishable to the naked eye for concentrations greater than 0.1 µM of lead. This suggests that the reporter signal becomes saturated at higher lead concentrations, making it difficult to visually differentiate between higher levels of induction.




Meanwhile, in the 500 µM and 1000 µM setups (as shown in Fig. 15), the colour intensity appeared to decrease. We hypothesize that this is due to the toxicity of these high lead concentrations, which likely exceeded the cells' resistance to lead poisoning. As a result, the cells may have experienced poor growth, leading to reduced expression of the reporter protein.

II. Cadmium Biosensor

We designed three cadmium biosensors, with the first utilizing a MerR repressor protein, which was originally a mercury-binding activator but was modified by the 2015 iGEM Team SCUT to specifically bind cadmium and repress the expression of pCadA. This modified MerR repressor is coupled with a pCadA promoter (BBa_K5152006), which contains the modified MerR binding site, to drive the expression of the amilCP reporter protein.

When cadmium binds to the modified MerR, it releases its repression on the pCadA promoter, allowing the expression of the downstream reporter protein, resulting in a visible blue colouration.

The functional study successfully demonstrated that this cadmium biosensor works as expected. We repeated the experiment three times, cells treated with 200 µM cadmium chloride after 12 hours of incubation already showed a blue colouration, while the control cells remained white (as shown in the figure below).

However, it's important to note that in one of the repeats, even though the cells were harvested in under 18 hours, both the control and experimental setups showed blue colouration, although the experimental setup exhibited a higher intensity of blue. This indicates that the biosensor exhibits leaky expression, a property that we will need to address in future studies.



Figure 17:

(a): A mild blue colouration from the amilCP protein is observed in the cadmium biosensor culture when 200 µM cadmium is added.

(b): In the pellet form, an obvious blue colouration from the amilCP protein is observed in the pellet of the biosensor culture with 200 µM cadmium added, confirming the activation of the cadmium biosensor.




In contrast, the other design utilizing the pYodA promoter (BBa_K5152007), which is reported to be cadmium-specific in E. coli, demonstrated a severe leaky expression issue. In both of our experimental repeats, while the cadmium-added setups consistently showed a deeper blue colouration, the negative control setups (without cadmium) also exhibited blue colouration (as shown in the figure below).

This significant leaky expression makes pYodA unsuitable as a promoter target for an effective biosensor, as it compromises the sensor's specificity and reliability.



Figure 18: Leaky expression of the amilCP blue protein was observed in the pYodA-containing cells, even when cadmium was not added. This indicates that the pYodA promoter exhibits significant leaky expression, making it less suitable for use in a cadmium biosensor.




The final design involved a modified pCadA promoter with ArsR and CrzA binding sites added upstream (BBa_K5152005), a design originally developed by the 2009 iGEM Team Newcastle. Unfortunately, we were unable to successfully assemble this construct. Despite attempting the cloning process three times, no colonies formed after transformation.

At this point, we still cannot determine the exact reason for the failure, and further investigation will be needed to identify and resolve the issue.

III. Zinc Biosensor (BBa_K5152009)

The design of the zinc biosensor is relatively simple, as it utilizes the pZnt promoter, which is part of the native Zinc regulatory operon in E. coli. This system works with the native ZntR repressor protein found in E. coli; when zinc is present, it releases the inhibition of ZntR on the pZnt promoter, allowing the expression of the downstream reporter gene, eforRed.

However, similar to the issues observed with the pYodA cadmium biosensor, we consistently observed leaky expression in all experimental repeats. The biosensor without zinc added showed a mild red colouration from the eforRed protein, indicating that there is leaky expression even in the absence of zinc. Although the cells with zinc consistently showed a deeper red colour than the controls, the leaky expression makes this design less than ideal as a reliable biosensor, as it could lead to false positives or false alarms (as shown in the figure below).

Further investigation is needed to address this issue and reduce the leaky expression, possibly by optimizing the promoter or regulatory elements in the system.





Figure 19: Leaky expression of the eforRed red protein was observed in the pZnt-containing cells, even when zinc was not added. This indicates that the pZnt promoter is prone to leaky expression, making it less reliable as a biosensor for zinc detection due to the potential for false positives.

IV. Mercury Biosensor (BBa_K5152008)

For this design, we made a major mistake, and we realized it when we observed that one of the colonies formed after transformation was bright pink in colour. Initially, we thought the pink colour was due to contamination from other plates, but after consistently seeing more red colonies forming, we decided to go back and check the DNA.

Upon review, we discovered that we had accidentally entered the sequence of the lead-sensing construct when ordering the DNA for the mercury biosensor construct from IDT (as shown in the pictures below). Unfortunately, by the time we identified this mistake, there was no time left to re-order the correct DNA sequence.

As a result, there are no further results for this biosensor.



Figure 20: Our careless mistake where we accidentally copied the same sequence of the PbrR construct (lead biosensor) when placing the order for the pMerT construct (mercury biosensor). This error resulted in the wrong DNA being synthesized, and as a result, we were unable to proceed with the experiments for the mercury biosensor.




However, before finding out the mistake, we performed the functional study on the incorrect colonies. Interestingly, we found that in multiple repeats, the cultures to which we added 100 µM of mercuric chloride showed no cell growth (as shown in the figure below). This is highly likely due to the extreme toxicity of mercury, which killed all the cells at this concentration.

This finding suggests that 100 µM of mercury is too high a concentration for such studies, and we must be very careful in future investigations involving mercury. Lower concentrations should be used to prevent complete cell death and to allow us to accurately assess the biosensor's response.



Figure 21: No growth of E. coli was observed in the culture after 16 hours of incubation. The supposed "MerR-pMerT" construct we were testing was actually the PbrR-pPbr lead-sensing construct, as we mistakenly ordered the wrong DNA sequence. The lack of growth is likely due to the high toxicity of 100 µM mercuric chloride, which killed all the cells.

8. Learn

Through the experimental engineering cycles, we gained valuable insights into our project. We learned that employing an effective and systematic problem-solving approach was key to successfully cloning our constructs and conducting investigations on our designs. Each iteration of the cycle allowed us to refine our methods, troubleshoot issues, and adapt our experiments based on the challenges we faced, ultimately leading to a deeper understanding of both the engineering process and the functionality of our biosensors.

Conclusions of our results

1. Validation of Chromoprotein Expression: We successfully validated the expression of the chromoproteins and integrated them into new composite designs, showcasing their characteristics. However, we observed that the expression of cjBlue in E. coli seemed to cause difficulties in expression. Further investigation is needed to determine the underlying reasons for this issue.
2. Lead Biosensor Success: Our lead biosensor effectively detected the presence of lead and initiated the expression of chromoprotein as a reporter gene. The expression of the chromoprotein produced a visible signal that could be observed with the naked eye, without requiring any specialized equipment or techniques.
3. Concentration-Dependent Lead Detection: The lead biosensor demonstrated concentration-dependent behavior, where the intensity of the colour correlated with the concentration of lead within a certain range. This feature enhances the versatility and practicality of the biosensor, making it more useful for various applications.
4. Cadmium Biosensor Success: Our cadmium biosensor design (MerR-pCadA) was also able to detect the presence of cadmium, showing functional expression and detection capabilities.
5. Leaky Expression in Other Biosensors: Our other biosensor constructs, including pYodA and pZnt, showed some potential in detecting their respective heavy metals. However, due to leaky expression, we cannot yet draw definitive conclusions about their reliability. Further repeats and investigations are required to confirm their validity and reliability.
6. Lessons from DNA Ordering: A critical lesson we learned is the importance of carefully checking DNA sequences before placing an order. Our experience with ordering the wrong sequence for the mercury biosensor highlighted the need for more rigorous checks to avoid such errors in the future.

With these results and conclusions, we have successfully demonstrated that our concept is viable and proven that our biosensor designs can work as intended. However, we are also fully aware that there is still significant room for improvement and that further investigations are necessary to refine the system.

Limitations and room for improvements of the studies

There are several major considerations that we need to address in our project:

1. Increasing the sensitivity to low concentrations: Sensitivity is crucial for a biosensor. The acceptable lead level in drinking water is around 0.03 µM (5 ppb by FDA and/or 10 µg/L by Hong Kong Water Service Department). In our results, the 0.01 µM Pb concentration showed no significant difference from the control. Therefore, we need to investigate ways to enhance sensitivity to detect lead at these lower concentrations.
2. Find out the sensitivity to different concentrations: We aim to decrease the interval between tested concentrations to determine how precisely the biosensor responds to small changes in metal concentration. This will help us assess the sensitivity and accuracy of the biosensor across a broader range of concentrations.
3. Integrate the biosensor into hardware and real-world applications: To use our biosensor in the automated warning and growing device developed by the hardware team, we need time to test its compatibility and use bacterial results to train the AI model for colour recognition. This integration will require a significant amount of time for testing and optimization.
4. Minimizing leaky expression in biosensors: A common issue is that reporter genes may show leaky expression due to culturing conditions, like prolonged incubation. In some designs, such as pZnt and pYodA, this occurs consistently in negative controls. Enhancing the expression of regulatory proteins can help tighten control, but it might reduce biosensor sensitivity. Careful further investigation is needed to address this problem.
5. Testing on solid medium / cell-free expression systems: We aim to test if the biosensor functions effectively on solid media or in a cell-free system. Success in these systems would significantly increase the applicability of the biosensor, offering more flexible and safer usage options.
6. Validate the other biosensor designs and assess combined use: We plan to continue investigating the biosensor designs we couldn’t complete, as our goal is to develop sensors for the four most common heavy metal pollutants. Additionally, we are interested in testing what happens when different biosensors are mixed in the same culture to evaluate their combined response to heavy metals.
7. Ensuring safety measures: Safety is critical to our project. We need to implement measures like a kill switch or containment mechanisms when working with engineered organisms. Additionally, exploring cell-free expression systems could help mitigate concerns regarding the use of genetically modified organisms.

Therefore, we're beginning a new round of the engineering cycle to address these important points. By combining our biological system with the device, we aim to improve our product's functionality, applicability, and safety. It's crucial for us to thoroughly investigate every aspect to ensure that our system meets the necessary standards and regulations for real-world use.