Achievements

---

We successfully deleted the csg operons from the genome of MG1655.

We successfully overproduced Curli fibers.

We successfully established a protocol to quantify captured microplastics.

We successfully captured microplastics.

We successfully cloned the Rubber-binding domain.

We designed the addition of Plastic-binding peptides and started cloning them.

Abstract

---

The aim of our project was to create an engineered biofilm, based on overproducing Curli fibers, to be able to capture microplastics. In this section, we summarized our results and the problems we encountered along the way of our project. It was not always a linear path, as we encountered several problems, but we always created new solutions and new ways to solve these problems.

Gene deletion

---

The first step in realizing our goal was to create a working strain. Our working strain had to possess a deletion of the csg operons, the operons that allows the production of biofilms, so that we could test the functioning of our plasmid. We decided to perform the csg operons deletion on E. coli strain MG1655. We chose the technique of red lambda deletion because this method allows precise insertion or deletion of a sequence inside the E. coli genome. We used three principal plasmids for performing lambda-red gene deletion. The first plasmid, pKD13, contained the FRT regions flanking a kanamycin resistance cassette that was then inserted into the genome in place of our operons. The second plasmid, pDK46, contained the red recombinase genes that allowed recombination between our resistance cassette and the csg operon. The red recombinases need to be induced by the addition of L-arabinose. The third plasmid, pCP20, encoding for the flippase enzyme that removed our resistance cassette via recombineering of the FRT regions (see section Engineering). The first step of this recombination process was to amplify the fragment of the pKD13 plasmid containing the kanamycin resistance cassette and the two FRT sites using PCR. By adding them to the primers, we incorporated the two homologous regions during the PCR amplification. We then checked the size of our fragment on a gel electrophoresis. We expected a size of 1330 bp and indeed, we obtained a fragment of this size.

We then transformed the pKD46 plasmid into our MG1655 strain, and selected the cells containing pKD46 by the ampicillin resistance. We then proceeded with recombination by transforming the fragment containing the kanamycin resistance cassette into strain MG1655 with the pKD46 plasmid. We induced the expression of the Red genes with L-arabinose and allowed the recombination to occur. To verify that the recombination had taken place correctly, we first performed colony PCR on four colonies, amplifying the region of the genome that contained the csg operons to check whether the resistance cassette had been inserted. We could see on the electrophoresis gel (Figure 2) that the fragment size corresponded to what we expected and that the resistance cassette had therefore been inserted.

To be sure, we still decided to send the DNA fragment for sequencing. Sequencing (Figure 3) showed us that both the kanamycin resistance cassette and the two FRT regions had been inserted.

Once we were sure that we had the correct insertion of the kanamycin resistance cassette, we started the procedure for its removal from the genome, so that we would have the strain without a csg operon and without kanamycin resistance that would prevent us from using plasmids with kanamycin resistance in the future. To do this, we inserted the pCP20 plasmid, coding for a flippase, into the strain containing the kanamycin resistance. To check the success of this step, we streaked single colonies on one plate containing kanamycin and one LB only. This allowed us to check which colonies were growing on kanamycin and which were not, so we could discard those that still had kanamycin resistance. We isolated a strain that had the csg operons deleted and the kanamycin resistance cassette was removed. We named this strain PB_002. Since we successfully performed the deletion of csg operons, we decided to test the phenotype of MG1655 onto plates with Congo red (CR) and Coomassie blue (CB) to see if it produced biofilm. We used strain M037 (a strain from the lab of our PI also containing a csg deletion) as a negative control. We expected that the colonies producing Curli fibers will appear red and dry, while those not producing Curli will appear white or light pink and moist (Figure 4). The colony of our deletion strain was pink. This visual test allowed us to conclude that the gene deletion had worked correctly, and therefore we were able to create our own strain with the csg operons deleted that did not produce biofilm.

Conclusion

In conclusion, we succeeded in deleting the csg operons using the red lambda recombination method, thus obtaining our working strain. We confirmed this deletion by Sanger sequencing and phenotype analysis on the ability of our csg knock-out strain to form biofilm. The creation of our working strain was essential for the continuation of our project, in particular to proceed with the testing of our plasmids.

Capturing microplastics

Curli fiber overexpression

Here we describe how we overproduced Curli fibers and then could show that we can use the resulting biofilm to capture microplastics. This part was carried out while we were constructing our deletion strain and therefore used a strain from the lab of our PI (M037), also containing a deletion of the csg operons. We transformed it with pC3 - a plasmid overproducing the Curli genes (see section Design). We tested the phenotype of M037 and M037 pC3 onto plates with Congo Red and Coomassie Blue to see if it produced biofilm. The use of plates with Congo red (CR 40 µg/ml) allowed us to visualize the expression of Curli fibers. In the plates we add Coomassie blue (CB 1 µg/ml) to enhance the contrast of the Congo red staining when looking at the colonies with the light microscope. We expected M037, does not produce the Curli fiber, due to the deletion of csg operons in the genome; and M037 pC3 does produce the Curli fiber, thanks to the plasmid pC3 constitutively expressing the csg operons. We followed the Biofilm protocol (see section Experiments, Biofilm protocols). We prepared and inoculated two different types of plates: LB CR CB for M037 strain and LB CR CB + Amp for M037 pC3 strain. After 24 hours of incubation, we expected that the colonies producing Curli fibers will appear red and dry, while those not producing Curli will appear white or light pink and moist. The results we obtained show us that M037 strain, with the csg operons deletion cannot produce Curli fibers, as we can see the colony is pink (Figure 5, A). While M037 pC3 strain shows an overexpression of Curli fibers production, as we can see the colony is red (Figure 5, B).

Microplastics capturing capacity

Having a strain that produced biofilm, we wanted to begin to understand how we could study the interaction between biofilm and microplastics and how we could quantify how much microplastic could be captured. As a first test to quantify the mass of captured microplastics, we tried to reproduce an experiment found in an article: “Quantification of dry weight mass of microplastics” . We expected the microplastics to end up as pellets, we dried and we weighed it. But this did not allow us to determine how much microplastic had been captured by the biofilm. We therefore decided to try another way. The next idea was to cultivate the biofilm with microplastics in a multi-well plate. Hoping that, after 24 hours, we could see aggregates of microplastics at the bottom of the well. But also with this technique, we could not get satisfactory results.
Finally, by reading articles, we managed to find a method for quantifying the captured microplastics and also the biomass produced by the strains. The idea was to use a vacuum pump system and filters with pores of the size that would allow us to determine what we were interested in: the biomass produced in one case and the microplastics captured in the other. The idea of the first test is to use the vacuum system with a filter with a certain pore size that allows us to retain the microplastics and the biofilm that has been produced. The cells that have not produced biofilm will pass through the filter, allowing us to quantify the biofilm mass produced. The idea of the second test is almost the same, only in this case we initially want to have the microplastics captured by the biofilm as pellets, while the microplastic remaining free is present in the supernatant. We then lyse the cells so that only the captured microplastics remain attached to the filter. This allows us to quantify the ability of the biofilm to capture microplastics. We prepared two different samples, (M037 and M037 pC3) for the two tests: the “Biomass quantification of biofilm” and the “Quantification of microplastics captured by biofilm”. For each sample, we made three replicates. The first test allows us to quantify the biomass of biofilm produce by the two strains: M037, does not produce the Curli fiber, due to the deletion of csg operons in the genome; and M037 pC3 does produce the Curli fiber, thanks to the plasmid pC3 constitutively expressing the csg operons (Figure 6). The results prove the overexpression of Curli fibers produced by the pC3 plasmid.

The second test allows us to quantify the microplastics captured by biofilm produced by the strains. We used two types of microplastics for the quantification: PE (polyethylene) and PP (polypropylene). Figure 7 shows us that the overexpression of Curli fiber allows us to successfully capture more microplastics.

Functionalization of the Curli fibers with Plastic-binding peptides

---

In order to further improve the plastic binding capacity of the biofilm, we wanted to add Plastic-binding peptides to the csgA. To read all the details about our different attempts of doing this, check the page Engineering success . Here we summarize the most promising attempt. We identified Plastic-binding peptides from the literature. Our aim was to fuse them to csgA via a linker. We PCR-amplified fragments containing the linker and the specific peptides. We were expecting a size of 149 bp, which is what we obtained (see Figure 8).

By PCR we also amplified the plasmid pC3. Next, we tried to assemble the two fragments using Gibson Assembly. After transformation, we checked by Colony PCR and obtained fragments that correspond to the expected size (1036 bp) (Figure 9).

We then sent these PCRs for sequencing.

Unfortunately, the linker and specific peptides were not present in the plasmid (see Figure 10). We suspected that the primers we used amplified the csgA present in NEB5α cells and that we sequenced the genome region. Therefore, we isolated the cloned plasmids and sent the plasmids plasmid for sequencing. The results confirmed that the peptide and the linker were not present.

Conclusions and outlook

---

We hypothesize that the colonies that grow after transformation contain the original pC3 that we used as template during PCR. As the modification is only 149 bp it is impossible to differentiate in our gel the two plasmids. To prevent this from happening again, we should digest our PCRs with Dpn1, an enzyme that digests methylated DNA. The original plasmid in NEB5α is methylated and would be digested, leaving only the amplified non-methylated DNA. Thus, we are sure that all transformants come from the Gibson Assembly . After inserting the peptide sequences successfully into the plasmid we would then next test whether their catching to the specific microplastic capacity is greater than with the Curli fibers alone.

Functionalization of the Curli fibers with a Rubber-binding domain

---

In parallel to design the plastic specific peptides we decided to extend our project and find a way to also catch the rubber particles from car tyres. In literature, we found a protein domain (Als3) that has been shown to bind to rubber. ² As the rubber domain is too big to be directly fused to CsgA and exported, we used a different approach: we used a system composed of two parts, that would interact via the SpyTag/SpyCatcher system.
So we created a new composite part, in which the Rubber-binding protein is fused to the SpyCatcher and this would interact with the second part by forming a covalent bond to the SpyTag domain fused to csgA. To achieve this, we used two different plasmids: one plasmid containing the csg genes including csgA fused to the SpyTag, and the second plasmid containing the genes for the Rubber-binding protein fused to the SpyCatcher (Figure 11).

The plasmid pFM_tag is identical to the plasmid pC3, with the only difference that the csgA gene fused to the SpyTag sequence. This plasmid was kindly provided by one of our assistants (Roberto Avendaño Vega). The plasmid that contains the genes for the Rubber-binding protein fused to the SpyCatcher was engineered with success by our team. (see section Engineering - Rubber-binding domain)
Once we obtained the correct plasmid for the synthesis of the Rubber-binding domain, that is specifically constructed for high-level expression of recombinant proteins. Our working strain PB_002 (strain lacking the csg operon) will instead express the Curli fibers fused to the SpyTag.

Binding between the two parts

To test if the interaction SpyTag/SpyCatcher worked correctly we lysed the cells and run a SDS-PAGE (see section Engineering). The three overexpressed protein complexes (CsgA-SpyTag, Rubber-binding domain-SpyCatcher and the complex of the two) differ in their molecular weight so we would expect to see three different bands that would stand out from the background of the cellular proteins. We decided to run the first SDS-PAGE directly on the crude cell lysates, hoping that the proteins would be overexpressed enough to see them clearly in the gel. Unfortunately, we see in Figure 13 it was not the case. There are bands in the expected regions, but we can be sure that they correspond to the proteins that we want to identify.

Faced with these results, we decided to redo the process, but adding the purification step with the His-tag method. As the Rubber-binding domain contains a His-tag, this purification process would enrich the Rubber-binding domain and the Rubber-binding domain linked to the CsgA through the SpyTag-SpyCatcher interaction for the mixture of the two strains. This process would allow us to hopefully have more clean bands.

From the gel, we still cannot identify clear bands for the proteins we are looking for. We think this result is due to an error made during the His-tag purification.

Conclusions and Outlook

From the results that we have obtained until now it is possible to conclude that we were able to correctly assemble the plasmid that contains the Rubber-binding domain fused to SpyCatcher (as shown by Colony PCR and sequencing). Future experiments would repeat the His-tag purification. Next, we would test the effectiveness of this mechanism within a solution with rubber particles. In fact, it would be necessary to adapt the microplastic quantification test with the vacuum flask to the rubber particles and the two-plasmid system. We would then compare the quantity of rubber particles captured by the wild type Curli fibers and the modified Curli fibers with the Rubber-binding system.

Plasmid engineering

---

IPTG-inducible plasmid for Curli fibers

During our project, we also wanted to develop a plasmid that would allow us to control the overexpression of Curli fibers (the main protein component of the biofilm matrix) using Isopropyl β-d-1-thiogalactopyranoside (IPTG). After obtaining the plasmid, we wanted to transform it into our MG1655 Δcsg strain so that the expression of biofilm will be conditional only in the presence of IPTG induction. We decided to create the plasmid by modifying an already existing plasmid, called pFM_1300.¹ This plasmid encodes the csg operons (containing csgABC and csgDEF genes), with promoters that respond to light and IPTG, respectively. There is also the lacI sequence, necessary for IPTG induction. We wanted to replace the light-inducible promoter with an IPTG-inducible promoter.

To build the plasmid, we used the cloning protocol (see section Experiments). The template for the construction of pFM_indu_Ctrl was pFM_1300. The first step of the process was to do two PCRs on pFM_1300 to obtain two fragments. The first one would contain the Plac promoter followed by the gene cgsABC and the gene for the ampicillin resistance (fragments size: 3884 bp). The second fragment instead would contain the genes cgsEFG (fragment size: 4384 bp) To confirm the correct amplification of the two fragments necessary for assembling pFM_indu_Ctrl, we run an electrophoresis (See Figure 16). We successfully obtained the correct fragments since their size corresponds to the expected sizes.

After obtaining the correct fragments, we performed a Gibson Assembly to create the pFM_indu_Ctrl plasmid. Then, we transformed the Gibson Assembly mix by electroporation into E. coli NEB5ɑ and plated the cells to obtain individual transformants. As the last step, we did a Colony PCR to select for a transformant with correctly assembled plasmid. Unfortunately with the Colony PCR, we did not obtain the results that we were hoping for. For the first junction (region after the csgG gene) we have obtained the expected length, but for the second junction (region that contains the Plac) we have obtained two bands instead of one (see Figure 17).

We hypothesized that the problem was caused by the primers. There was a possibility that the primers that were responsible to amplify the junction in the pLac region, were capable of attaching to multiple sites on the plasmid or even to the genome of E. coli strain NEB5ɑ. We decided to redesign a new couple of sequencing primers. We also decided that to identify the correct plasmid, we would extract the plasmid from the colonies with a Miniprep instead of doing a Colony PCR and then directly send for Sanger sequencing with the new primers. In this way, we could avoid possible attachments of the primers to the E. coli genome.
This time, we received a positive result from one of the candidates that we sent for sequencing (see Figure 18), indicating that we successfully inserted the Plac promoter.

We also sent our plasmid for full plasmid sequencing to be completely sure that our plasmid is correct and contains all the necessary genes. Unfortunately, from these sequencing results it seemed that the genes csgB, csgA and lacI were missing. Our engineered plasmid pFM_indu_Ctrl seemed to be incapable of producing biofilm since the genes coding for the basic unit of Curli fibers are missing.
While we were waiting for the sequencing results, we decided to start testing for biofilm formation. We wanted to know whether the plasmid was IPTG inducible and whether it could produce biofilm. We used 'Qualitative characterisation of biofilm using Congo red' protocol (put link for experiments). The use of plates with Congo red (CR 40 µg/ml) allowed us to visualize the Curli fibers, a characteristic of biofilm formation. In the plates we add Coomassie blue (CB 1 µg/ml) to enhance the contrast of the Congo red staining when looking at the colonies at the light microscope. For this experiment, we included the deletion strain PB_001 (MG1655 Δcsg::kan strain), and strain PB_001 with the pFM_indu_Ctrl plasmid. After 24 hours, we expected that the colonies producing Curli fibers will appear red and dry, while those not producing Curli will appear pink and moist. The results we obtained showed us that the pFM_indu_Ctrl plasmid enabled biofilm production.

Conclusions

In this section we have described the construction of the plasmid pFM_indu_Ctrl. From the results it seems that the cells when transformed with our plasmid are capable of producing biofilm. Unfortunately the result of full plasmid sequencing of pFM_indu_Ctrl does not support this statement, since it has shown a lack of the fundamental genes for the biofilm formation. After lengthy discussions with assistants, we came to the conclusion that we were probably not working with a single plasmid all this time, but with a mix of plasmids of different sizes and composition. Probably during the Gibson Assembly, several plasmid versions were assembled and transformed into the same colony. This hypothesis might explain the fact that despite the sequencing result, the cells are able to form the biofilm. Faced with this, we decided not to invest more time into getting a clean version of pFM_indu_Ctrl, but continue with plasmid pC3 instead. This did not stop us from continuing our project and finding a way to demonstrate that the biofilm is capable of capturing microplastics and to make it specific for capturing certain types of plastics through the addition of specific peptides. The next phases of the project were based exclusively on the pC3 plasmid.

Constitutive Biofilm Production and Negative control

---

One of our objectives at the beginning of the project was the designing of two other plasmids: pFM_pos_Ctrl and pFM_neg_Ctrl (Figure 20). The pFM_pos_Ctrl would allow the production of Curli fibers constitutively, while pFM_neg_Ctrl would lack one of the necessary genes for the Curli production. For both plasmids we used as a template pFM_1300. To build the pFM_pos_Ctrl we have decided to replace the Plac promoter with two different constitutive promoters, which sequence was taken from the iGEM database (Bba_23100 for csgBAC and Bba_23119 for csgDEF)

Instead for the construction of the pFM_neg_Ctrl we have not only replaced the pLac promoters with constitutive ones, but we have also deleted the gene csgA, so it will not produce the Curli fibers. To build the two plasmids we have followed the Cloning protocol (see section Experiments). Unfortunately, we were not able to clone these two plasmids. We faced several challenges as documented in our notebook. One of the problem was that the primers we used, also bound to the csg operon of our cloning strain NEB5α. A potential solution would be to carry out the cloning in our working strain with the csg operon deleted. However, as these plasmids were not absolutely essential to continue our project, we decided to not invest more time in creating them and instead focused on demonstrating the plastic capturing capability of our biofilm.