Introduction

With our current years project, we tried to achieve two goals: bacterial chemotaxis-driven detection of Cadmium as well as Cadmium Bioremediation through increased uptake and sequestration within the cytoplasm of engineered E. coli strains. To this end, we conducted several gain-of function studies to test both the feasibility of our concepts as well as the functionality of their individual components.

In order to generate a bacterial strain capable of Cadmium detection, we introduced a modified methyl-accepting chemotaxis protein (MCP) into an otherwise chemotaxis deficient strain (E. coli UU3330, kindly provided by Prof. Parkinson, University of Utah). This modified periplasmatic MCP carries several point mutations that reform its binding pocket to accept Cadmium instead of Ribose as its ligand. For this Cadmium-specific Ribose binding protein (CdRBP1) to induce flagellum-driven swimming behavior in our chemotaxis-deficient strain, we had to co-express its respective inner membrane receptor, Tsr, to transmit signals across the inner membrane. Furthermore, we had to simultaneously re-introduce a knocked-out chemotaxis receptor, Trg, which is hypothesized to form a complex with Tsr ensuring its proper function.

Testing our chemotaxis-driven Cadmium detection concept, we planned on introducing our modified strains onto LB-agar plates as part of a drop plate assay. As co-expression of all our components is necessary for our desired phenotype, we were not able to individual functionality through our gain-of function study. Instead, also due to the limited timeframe of our project, we went ahead and tried confirming Cadmium-specific swarming behavior directly.

As for our Bioremediation concept, we again relied on the co-expression of several exogenous as well as endogenous proteins in E. coli. Increased uptake of Cadmium was supposed to be achieved through the over-production of MntH, an endogenous Mangane-importer shown to also transport other bivalent anions. Imported Cadmium would be sequestered through either the small cysteine rich protein hMT2 or the products of a the phytochelatines synthetase AtPCS1. As both storage solutions required many cysteines to be incorporated into proteins, we hypothesized that the co-expression of a sulfate importer CysP together with a modified cysteine-synthetase CysE* would support or Bioremediation approach.

To test the components of our Bioremediation solution, we again relied on several gain of function studies to show the functionality of our constructs. Here, we first set out to test the uptake of sulphate in liquid E. coli cultures transformed with our gene cassette, before moving onto testing cysteine synthesis and cadmium uptake in the same manner. Finally, we wanted to investigate whether the expression of EcCARs, a cysteine-tRNA synthetase with the property to form polysulfides on proteogenic cysteines, would boost the uptake of Cadmium as well as lead to the production of Cadmium-sulfite quantum dots.

Chemotaxis

To obtain a proof of concept for our bioremediation approach, we had to show a specific chemotaxis-driven swarming behavior of a chemotaxis-deficient E. coli strain transformed with our Chemotaxis-construct. To this end, we conducted drop-plate assays testing the response of homogenously distributed bacteria towards different natural chemoattractant as well as several bionic anions to show specificity towards cadmium (Fig.1).

Analyzing the ring-formation around different compounds on our plate reveals several problems when it comes to proving our chemotaxis-based Cadmium detection concept. Ring formation around the area where a compound was added to the Bacteria-containing agar is indicative of chemoattraction towards the respective compounds. While we can show a desired ring formation of our transformed chemotaxis-deficient E. coli strain (UU3330_CHIU), the same ring formation is also visible on plate containing the UU3330-WT as well as our chemotaxis-capable strain E. coli BL21-WT. Given that Cadmium usually acts as a chemorepellent due to its toxicity, one could assume that the observed discoloration on each plate is indicative of dead E. coli instead of chemotaxis-driven behavior. Furthermore, we can also see ring-formation around almost all bivalent anions administered to the drop-plate assay as a control regarding Cadmium-detection specificity. Again, as this pattern can be seen both for our untransformed UU3330-WT strain as well as the BL21-WT control, we have to assume that this observable is not related to the introduction of our Chemotaxis construct. In summary, we were not able to prove the functionality of our chemotaxis-based cadmium detection approach with our drop-plate assay.

Bioremediation

To obtain a proof of concept for our Cadmium-bioremediation approach, we set out to simultaneously test the functionality of the different systematic components we have included on our Bioremediation construct. Here, we aimed at showing both increased uptake of sulphate and cadmium as well as cysteine synthesis of transformed E. coli in liquid culture. First however, we wanted to test whether the expression of our storage proteins alone as well as together with the other proteins on our bioremediation cassette could improve E. coli’s survival rate under increasing Cadmium concentrations. To this end, we performed a dot blot survival assay (Fig.2).

While the majority of E. coli cultures transformed with different construct combinations do not show better growth than their respective untransformed Wild-type, we are able to discern slightly better growth behavior for Bl21 solely transformed with EcCARs, hMT2 and AtPCS1 respectively. While we are not able to show this effect being translated into hosts co-expressing theses Cadmium storage solutions together with other parts of our Bioremediation cassette, we went ahead to analyze the constructs functionality further.

To prove the supportive role of CysP within our Bioremediation expression system, we performed sulphate uptake assays. Here, liquid cultures expressing all parts of the Bioremediation cassette alone as well as together in different combinations were analyzed towards the concentration of sulphate within the surrounding medium over time (Fig. 3).

While we are able to show a decrease in the mediums sulphate concentrations in the medium of E. coli expressing CysP both alone as well as together with CysE, it is not discernibly increased in comparison to the wildtype as well as E. coli expressing CysP alone. We therefore were not able to show proof for an increased sulphate uptake in strains expressing our sulphate importer.

To show that the individual expression of constructs from our Bioremediation cassette had the intended effect on our host, we had planned to also assess cysteine synthesis in E. coli expressing our modified CysE variant. However, due to time issues, we instead decided to directly test the Cadmium uptake of our entire Bioremediation system. We were hoping to show that co-expression of CysE with our Cadmium importer as well as storage proteins would show increased amount of sequestered Cadmium when compared to strains expressing those components alone.

To quantify the uptake of Cadmium into different transformed E. coli, we measured its concentration in the medium of liquid bacterial cultures over time (Fig. 4). Unfortunately, we were not able to show consistently increased uptake of Cadmium for E. coli transformed with our storage proteins AtPCS1 and hMT2, both when expressed alone as well as together with the other components of our Bioremediation cassette (Fig.4a). Strikingly, the expression of our Cadmium importer MntH does not seem to affect this trend. We are however able to discern that the overproduction of EcCARs alone as well as together with CysP and CysE does seem to induce increased and consistent sequestration of Cadmium (Fig. 4. b.). Given the properties of the EcCARs mediated formation of cadmium sulfate quantum dots, we decided to follow up on this result.

It is known that complexation of cadmium in the cytoplasm yields quantum dots made of cadmium sulphate, the production of which can also be enhanced by different strategies in our case incorporating EcCARs together with the cofactor PLP in our constructs. Unfortunately, the deadline of the project cut our investigation short, but we were able to briefly compare the fluorescence of the E.coli pellets obtained from our uptake liquid cultures under UV light exposure (Fig. 5 & 6).

We can observe in Figure 5 that the blue colour stems from the intake of cadmium ions from the environment and the formation of quantum dots. While not being able to analyse the absorption and emission spectrum of each individual probe, we took photos of the resuspended pellets, that can be seen in Figure 6.

No major trends were observed during the short analysis and due to the wildtype showing roughly similar levels of quantum dot formation and poor comparability of the recorded images themselves, there will be no further analysis of the quantum dot production, which would need to be inquired more extensively to draw conclusions about the different constructs.

Conclusion and outlook

In summary, we were not able to deliver an unambiguous proof of concept for neither our chemotaxis-based detection nor our bioremediation approach. An underlying cause may be the fact that an interactive building and testing approach were both expression and function for each of the different components of our systems was not performed. Instead, we had to rely on an overall gain-of function proof given the relatively short time we were able to spend in the laboratory. This posed as a risk, given that the intended function of several of our co-expressed proteins was highly questionable without proper characterization and fine-tuning of expression levels. Without further testing we are not able to explain or propose a theory of why we were not able to achieve our desired phenotypes.

Future efforts should initially focus on a re-evaluation of our cloning approach. We have seen that the use of constitutive promoters has been problematic when it comes to plasmid amplification in our expression strains. The targeted introduction of inducible promoters especially in expression cassettes encoding for membrane proteins such as Trg, Tsr and MntH might alleviate those problems and allow to control the burden on E. coli when it comes to the overexpression of several proteins at once, as necessary for our both our systems. Furthermore, his-tagged variants of each of our individual constructs could help to quantify expression levels of individual components when introduced as Level1 constructs.