Our project aimed to create a bioremediation tool to purify mercury-contaminated waters using cyanobacteria. The design of our detoxification system is based on two genes, merA and merB, which code for mercury degrading enzymes. First we aimed to assemble plasmid constructs with these genes of interest that we could then transform to Escherichia coli (E. coli). After confirming the success of the assembly, we would transform them to a naturally competent genus of cyanobacteria, Synechocystis sp. PCC 6803. The modified Synechocystis sp. PCC 6803 strains would then be tested for their ability to resist and detoxify organic mercury compounds. This text will discuss our lab results and achievements.

Submitting the Project safety form

After careful evaluation of the safety protocols needed for our project and that our activities were compatible with iGEM regulations, we filled the Project safety form which was submitted by our Principal Investigator.

Acquiring a lab space

Our team acquired a lab space from the Department of Life Technologies in the University of Turku. We went through an orientation and safety training arranged by the department before beginning lab work.

Selecting the target genes for the project

As candidate expression targets for the project, we chose to use two merA genes: one native gene from Synechocystis and the other from Pseudomonas aeruginosa. As merB genes we chose to use gene homologs from E. coli and P. aeruginosa.

Choosing the expression host

We selected the photoautotrophic model organism, a genus of cyanobacteria, Synechocystis sp. PCC 6803 as the expression host for our project since we wanted to create a sustainable bioremediation tool that can produce its own organic compounds using light and carbon dioxide readily available for large-scale systems.

Selecting the construct assembly platform

An experimental Golden Gate based assembly system developed specifically for translation-level optimization of heterologous genes and operons in Synechocystis sp. PCC 6803 was selected for building the expression constructs.

Designing the expression system

An autonomous replicative expression vector pDF with an IPTG-inducible lac-promoter variant was selected as the expression construct. The specific translational control elements, ribosomal binding sites (RBS), to be fused with merA and merB were selected based on the article by Thiel et al. (2018). We used three different RBS’s (RBS S3, RBS S4 and RBS S5) for P2 and P4 positions in our plasmids. We designed a plasmid construct containing seven fragments: P0 pDF backbone, P1 promoter, P2 RBS, P3 merA, P4 RBS, P5 merB and P6 transcription terminator (Fig. 1). In the control assemblies, we used reporter genes EFE and sYFP2 instead of the merA and merB genes and we also made some assemblies with merA or merB with EFE or sYFP2.

Figure 1. Plasmid maps for our constructs using P2 RBS S3 and P4 RBS S5 for our RBS’s: A) sYFP2/EFE B) merA/EFE C) sYFP2/merB D) merA/merB

Modifying the target genes for the assembly

The target gene sequences were codon harmonized for Synechocystis sp. PCC 6803. We also modified the gene sequences to be compatible with the modified Golden Gate assembly system by removing internal BbsI restriction sites and introducing site-specific flanks.The sequences were also checked for no BsaI or SapI restriction sites to be compatible with iGEM systems.

Confirming the design by in silico assembly

The intercompatibility of the designed genetic fragments was verified in silico by assembling the fragments into expression plasmids in different combinations (Fig. 2). The assembly was tested with the different RBS elements in one-gene and two-gene constructs using merA, merB, and a control reporter gene sYFP2.

Figure 2. The expression plasmid was assembled in silico using SnapGene. The backbone (pDF) is cut out of the figure. Our target genes were designed to fit into positions 3 and 5, with the promoter in position 1, RBS’s in positions 2 and 4, and the transcription terminator in position 6.

Preparing the genetic element library

We completed a part library consisting of all the genetic elements (apart from merA and merB) needed for the assembly by amplifying the respective pre-existing constructs in E. coli. The plasmid minipreps and the respective E. coli glycerol preps were indexed in an electronic library designated for our team, and stored at -80 °C for later use.

Ordering merA and merB as synthetic fragments

The designed merA genes (Synechocystis and P. aeruginosa) and merB genes (E. coli and P. aeruginosa) were first ordered as synthetic linear gene fragments.

Testing the assembly system using the ordered linear gene fragments

We first evaluated the construct assembly for merA-merB operons using ordered linear merA and merB fragments. Control reactions with plasmid-based reporter genes sYFP2 and EFE were successful, but the merA-merB reactions did not yield positive clones. (Fig. 3).

Figure 3. 1% standard agarose-gel analysis of the sYFP2/EFE, backbone and merA-merB constructs using linear gene fragments. Quick-Load 1 kb Extend DNA Ladder was used as a marker (sample M). For the RBS’s we used P2 RBS S3 and P4 RBS S5. A) Sample 1 contains a positive control with YFP/EFE, sample 2 contains only the backbone and sample 3 contains the negative control with water. Samples 4-14 are the sYFP2/EFE constructs and samples 15-22 are backbone samples. B) Sample 1 contains only backbone and sample 2 the negative control. Samples 3-19 are the merA-merB constructs and samples 20-23 the backbone.

Testing the assembly system using PCR-amplified linear gene fragments

To increase the concentration of the merA and merB fragments in the assembly, we amplified the genes with PCR using primers designed for this specific purpose. Both the control reactions with plasmid-based reporter genes and the merA-merB plasmid constructs were unsuccessful (Fig. 4).

Figure 4. 1% standard agarose-gel analysis of the merA-merB constructs with PCR-amplified linear genes. NEB Quick-Load 1 kb Extend DNA Ladder was used as a marker (sample M) in the A and B gels and NEB Quick-Load 1 kb Plus DNA Ladder in the C and D gels (sample M). A) Sample 1 was the positive control with YFP/EFE, sample 2 only the backbone and sample 3 a negative control with water. C) Sample 1 was a negative control with water, sample 2 only the backbone and sample 3 the positive control with YFP/EFE. Samples in the A and B gels are PCR-amplified linear genes and in the C and D gels samples were more concentrated before PCR. All of the samples except the controls contain merA-merB constructs with P2 RBS S3 and P4 RBS S5.

Ordering merA and merB as plasmid-based gene fragments

Since we could not produce the designed merA-merB plasmid constructs using linear gene fragments, we ordered the genes in plasmid vectors.

Testing the assembly with plasmid-based gene fragments

We performed the assembly reaction with different combinations of the ordered merA and merB genes that were already in plasmids. We did not get positive PCR results for the merA and merB containing plasmid clones although the transformation results were promising (Fig. 6, 7).

Figure 6. Transformation results of construct assemblies. The constructs contained: A) the pDF backgone B) the part of expression plasmids that do not contain the pDF backbone but an insert consisting of 9 fragments C) merA from Synechocystis and merB from E. coli D) merA from Synechocystis and EFE E) merB from E. coli and sYFP2 F) sYFP2 and EFE.

Figure 7. 1% standard agarose gel with plasmid-based gene fragments. NEB Quick-Load 1 kb Plus DNA Ladder was used as a marker (sample M). A) sample 1 was the positive control (YFP/EFE), sample 2 only the backbone and sample 3 the negative control with water. Samples 4-7 were merA(P)/EFE constructs with P2 RBS S3 and P4 RBS S4, samples 8-15 merA(S)/merB(P) constructs with P2 RBS S3 and P4 RBS S4 and samples 16-24 merA(S)/merB(P) constructs with P2 RBS S3 and P4 RBS S4. B) Samples 1-8 were merA(P)/merB(P) constructs with P2 RBS S3 and P4 RBS S4, samples 9-16 were merA(P)/merB(P) constructs with P2 RBS S4 and P4 RBS S3 and samples 17-24 were sYFP2/merB constructs with P2 RBS S3 and P4 RBS S3. C) Samples 1-8 were sYFP2/merB(P) constructs with P2 RBS S3 and P4 RBS S4, samples 9-16 were merA(S)/EFE constructs with P2 RBS S3 and P4 RBS S5 and samples 17-24 sYFP2/merB(P) constructs with P2 RBS S3 and P4 RBS S3. D) Sample 1 was only backbone, sample 2 a negative control with water and sample 3 the positive control with a 9-fragment construct. For the RBS’s we used P2 RBS S3 and P4 RBS S5. Samples 4-15 were merA(S)/merB(E) constructs and samples were 16-24 merA(S)/EFE. E) For the RBS’s we used P2 RBS S3 and P4 RBS S5. Samples 1-3 were merA(S)/merB(E) constructs, samples 4-14 were sYFP2/merB(E) and samples 15-24 were sYFP2/EFE constructs.

Planning future experiments

Confirming successful assembly of plasmid constructs

Although we did not have time to assemble working merA-merB plasmid constructs, we planned out our next experiments which we would have performed after successful assembly. First, we would have confirmed the obtained E. coli clones using colony PCR and SDS-PAGE analysis. Successful assembly of sYFP2-merA reporter constructs would have been verified by fluorescence analysis. After that, we would have deposited the verified E. coli clones in our library.

Transforming the verified expression constructs in Synechocystis sp. PCC 6803

After extracting the assembled plasmids from the E. coli clones, we would have transformed Synechocystis sp. PCC 6803 with the assembled plasmids containing the merA and/or merB genes. Using antibiotic selection, we would have chosen antibiotic-resistant colonies and later also confirmed that they contain the desired plasmid constructs. After the confirmation, we would have deposited the verified cyanobacteria expression strains in our library.

Analysis of the cyanobacterial expression strains

To test the function of the Synechocystis sp. PCC 6803 cell strains that had been transformed with merA and merB genes, our goal was to perform qualitative and quantitative analysis on the expression strains. We would have compared merA-merB constructs and constructs with only merA or merB. Different gene homologs and RBS would also have been compared. First, we would have performed methylmercury and mercury(II) chloride resistance experiments on the modified cyanobacteria to indirectly get information about the function of MerA and MerB proteins. We could also have used proteomics to show that the MerA and MerB proteins are present in the expression strains.

We also contacted SYKE (FInnish Environment Institute) and Ruokavirasto (Finnish Food Authority) to determine if quantitative mercury analysis using ICP-MS would be possible. With ICP-MS, we would have measured the changes in methylmercury and mercury ion concentrations in different cell cultures. This would have required preprocessing of the samples since ICP-MS only can detect the total concentration of mercury.

Later, the constructs could additionally be tested with real contaminated water to see how the expression strains function in a biologically relevant environment. After the analytical steps and comparison, the expression strains could have been optimized and then analyzed again. Our final implementation plan is to build a MercuLess-truck but before it is possible, many additional steps are needed after creating a functioning cyanobacteria expression strain.

Discussion

Despite careful construct design, the construct assembly turned out to be difficult to implement. Our transformation and antibiotic selection results of the last assembly indicated that the transformed colonies had plasmid constructs inside them but our colony-PCR results were inconclusive. Therefore, it is impossible to draw reliable conclusions of the last, and also most promising, assembly results. More time would be needed to perform the assembly reactions again and proceed to analytical steps in making a merA-merB expression strain of Synechocystis sp PCC 6803.

Since our transformation and antibiotic selection, as seen in the plates (Fig.6), were successful, we can safely presume that leading up to the colony-pcr all the reagents and conditions were accurate. We can see from the lack of colonies in the control plate B that the antibiotic selection worked and a couple colonies in the backbone plate A indicates that there is only little background. All of our agarose gel-electrophoresis results had a successful positive control which shows that our reaction setup and the reagents used were correct. Therefore the issue may be specific to e.g primers used in the colony-PCR or a part of the construct we were trying to assemble. More experimentation is needed before we can pinpoint the exact issue.

Since we did not have enough time to assemble and test the function of the desired merA-merB constructs, there are good opportunities for future research and iGEM teams to continue studying the opportunities of the use of mercury modifying genes in cyanobacteria. We believe that our cyanobacteria-based system is promising and since cyanobacteria are photoautotrophic, the system has potential to become a sustainable bioremediation tool needed for environmental protection.

Thiel, K., Mulaku, E., Dandapani, H., Nagy, C., Aro, E.-M., Kallio, P. Translation efficiency of heterologous proteins is significantly affected by the genetic context of RBS sequences in engineered cyanobacterium Synechocystis sp. PCC 6803. Microb Cell Fact 17, 34 (2018). https://doi.org/10.1186/s12934-018-0882-2

Table of contents