Our second aim of the project is to include a treatment strategy that should be activated upon detection of biofilm formation. Our treatment is based on the Dispersin B (BBa_K1659200) activity. Dispersin B releases cells from the biofilm via degradation of the polysaccharides, targeting β-1,6-glycosidic linkages in poly-N-acetylglucosamine (PGA) ^[1][2]. This is found in the biofilm of three major biofilm-forming bacteria E. coli, S. aureus, and S. epidermidis ^[1].

To deliver Dispersin B, we created a phagemid design by displaying it on the PIII protein of an M13 phage ^[3]. The exact design of the phagemid will be explained in Construct 2. The activity of Dispersin B was purposefully selected using the general infection and lysis function of the M13. This is because the M13 phage is specific to E.coli, but to be broadly applicable, Bye-O-Film needs to target other biofilm-forming bacteria as well. Using the other well-studied capacity of M13 phage-display we can display Dispersin B, utilizing its ability to degrade biofilm of not just E.coli, but also S.aureus and S.epidermidis.

Besides its broad applicability, the mode of Dispersin B operation includes another benefit. With the cleavage of β-1,6-glycosidic linkages, the bacteria in the biofilm are degraded to a planktonic form. Aiding the body to naturally get rid of the infection. In the case of antibiotic treatment, it also makes the bacteria more susceptible to antibiotics and greatly enhances the likelihood of successful treatment. Ultimately also leading to a reduction of the required antibiotics for sufficient treatment. 

With the CUG sensor as a basis, different reporter options were considered. We were informed about various options through our Human Practices endeavors. We looked into ideas such as measuring the redox potential of hydrogen peroxide by catalase activity, using the expression of streptavidin, and measuring the output via surface plasmon resonance. Lastly, we were also advised on bioluminescence and fluorescence as reporter signals. 

The catalase idea was soon disregarded as impractical, because of complications of measuring the redox potential in the complex environment of the body. The Surface Plasmon Resonance technique was abandoned due to price considerations. The other options, using luciferase and fluorescence, were especially attractive taking the recommendations as mentioned in the [AREA Framework] into account. Our efforts to use a luciferase cassette from plasmid pAKgfplux1 (Addgene Plasmid #14083) were postponed. The size of the cassette proved difficult during our cloning strategy. The applied strategy is further denoted in Construct 1C. 

Ultimately, the choice was made to utilize GFP (BBa_E0040). Not only is GFP much smaller, but it is also well characterized, has a stronger signal compared to luciferase, and is readily available. This design does create extra challenges for the electrical component. How will the light be excited? How will the sensor only pick up the correct emission wavelengths of GFP and RFP? These questions are answered on our [Electronics page]. 

Our next design consideration is, then, how to normalize the expression of GFP to a useful signal. Normally GFP emission is normalized by cell density (OD600). In our setup, where our biosensor should give output to an electronic sensor inside the body, it is not possible to apply this strategy. Therefore, we decided to express RFP (BBa_E1010) behind a constitutive PcI promoter (BBa_R0051). This not only allows us to normalize the GFP expression by the RFP expression, but it also allows us to assess the biosensor's viability by verifying the initial presence of RFP expression. To test the hypothesis that RFP could be used to normalize GFP expression, the inducible promoter Ptac (BBa_K180000) was put in front of the GFP. This construct was however not obtained due to recombination around the repeating terminator sequence, which resulted in the removal of the Pcl-induced RFP from the construct.

Another design consideration we tackled was the issue of varying promoter strengths between the constitutive promoter PcI and the FleQ repressed PpeI promoter. To bypass this problem, and to ensure both GFP and RFP are transcribed with the same promoter strength, the aforementioned PcI promoter was placed in front of the PcI controlled GFP. Putting both RFP and GFP under control of the constitutive PcI promoter, with the notable difference that GFP is also repressed by the cyclic-di-GMP dependent repression of FleQ, via the PpeI promoter region. 

Taking these considerations into account, we initially came to the following part (Figure 3): 

Figure 3: Plasmid map of initial Construct 1 (Part code: BBa_K4720015, Part composition: BBa_J34801 (RBS) + BBa_E0040 (GFP) + BBa_B0015 (Terminator sequence) + BBa_R0051 (PcI promoter) + BBa_B0034 (RBS) + BBa_E1010 (RFP) + BBa_B0015 (Terminator sequence)). Benchling link

The individual parts were obtained from previous iGEM distributions and the constructs were obtained by bio brick RFC[10] standard assembly.  This design incorporates RFP downstream of the constitutive promoter PcI, along with a Ribosome Binding Site (RBS). GFP, on the other hand, is positioned downstream of a distinct RBS. It is important to note that the design currently lacks the constitutive PcI and repressible PpeI promoters. Before introducing the PcI and PpeI promoters, two other improvements were desired. 

Envisioned improvements were the replacement of one of the double terminator (BBa_B0015) sequences with another terminator (BBa_B1006). By removing this sequence repetition, the construct should be less prone to recombination. Due to the time-consuming process of cloning and the need to switch cloning strategies, we were only able to create the following part, excluding the RFP expression system. 

Figure 4: Plasmid map of final progress made on Construct 1 (Part code: BBa_K472001, Part composition: BBa_E0040 (GFP) + BBa_B1006 (Terminator sequence)). Benchling link

When part BBa_K4720012 gets incorporated into the original BBa_K4720015, the envisioned theoretical part would be created:

Construct 1B_(1-3) Reporter control

Our final sensor product, Construct 4, ultimately uses the reporter function as described in Construct 1 and the sensor function as described in Construct 3. As earlier described, the expression of GFP will be under cyclic-di-GMP dependent FleQ control. A potential prospect of this final construct is that GFP will never be expressed. To check if it is possible to express GFP in our design, and to potentially pinpoint future problems to the repression and/or activation of FleQ, we created Construct 1B₁.

Construct 1B1 (BBa_K4720019), therefore resembles the promoter-deficient GFP intermediate from Construct 1. Construct 1B₁, in comparison, includes the IPTG inducible promoter Ptac (BBa_K180000). Firstly, with this construct, we can test if we can produce GFP in this system, without the PpeI promoter. Secondly, we can use this construct to measure the potential presence of Fluorescence Resonance Energy Transfer (FRET) effects. When taking the excitation and emission spectra into account for both GFP (BBa_E0040) and RFP (BBa_E1010), it is possible to see quenching of GFP in the presence of RFP. The opposite, quenching of RFP due to the presence of GFP should not be measured.

To measure this we created, besides Construct 1B₁, two other constructs. Construct 1B₂ contains an RFP behind a constitutive PcI promoter. Construct 1B3, our dual GFP and RFP expression system contains both fluorescent proteins. In principle, the expression of RFP, under the same conditions, should be the same between Constructs 1B2 and 1B3. As well as the expression of GFP when the same amount of IPTG is added to induce Ptac in 1B1 and 1B2. The difference in measured fluorescence between 1B3 - 1B1 shows the level of quenching of GFP by RFP, while the measured difference between 1B3 - 1B2 shows the level of quenching of RFP by GFP.

Construct 1C Reporter module (Luciferase - Theoretical)

The plasmid containing the luciferase cassette was obtained from addgene (pAKgfplux1). Because the cassette contains an XbaI restriction site, site-directed mutagenesis is required to make the part RFC[10] compatible. The primers, Figure 5 and Table 1, were designed with a 38 bp overlapping region for circularization, and were designed to amplify the new cassette for insertion into the pSB1A3 vector by Gibson Assembly.

Figure 5: Primers designed for site-directed mutagenesis to remove XbaI restriction site making the construct RFC[10] compatible.

After the Gibson assembly of this theoretical module, we could have combined Construct 1C (reporter module) with Construct 3 (sensor module), to generate the following part: 

The original design of the cyclic-di-GMP sensor was obtained from the 2022 CUG China team (BBa_K4242013). This design was ordered as synthetic DNA from TWIST bioscience in two parts. A codon optimization was performed on the FleQ coding sequence (BBa_K4720000). During the optimization, three SapI restriction sites were removed, making the part RFC[1000] compatible. The fragments were designed with 30 bp overlap at the ends for Gibson Assembly into the pSB1A3 vector, the final product is shown in Figure 7. The colonies transformed with the Gibson Assembly reaction mix were screened by colony PCR, without any result. In the samples, the FleQ coding sequence was consistently absent when reviewing the sequencing results.

A second attempt to obtain the cyclic-di-GMP sensor was made with SapI Golden Gate Assembly using the same codon-optimized fragment. The previously mentioned synthetic DNA fragments were amplified by Q5 PCR and the SapI recognition sites were added with the primer overhangs. The fragment sizes were confirmed on a 0.9% agarose gel. The colonies obtained from the transformation with the Golden Gate reaction mix were screened by colony PCR and restriction digests. None of the colonies contained the desired construct. After sequencing it was concluded that some samples contained empty backbones, while others contained only the promoter insert. The possibility of just the insertion of the promoter in the ligation with pSB1A3 is significant. Only one of the bases of the sticky ends is not matching (Figure 8). An explanation could be that FleQ circularizes because there is one mismatching base. After consultation with our supervisor, Julia, it was concluded that a possible explanation could be that the concentration of FleQ is lowered due to the circularization. To test this hypothesis a second attempt with the SapI Golden Gate strategy was carried out, but instead of equal molar ratios, the insert-to-backbone ratio was now tried at both 5:1 and 10:1. As the second attempt was unsuccessful, we sought the help of another supervisor, Patricia. During this meeting, it was hypothesized that the repeating sequence of the PcI promoter (which is both in front of the FleQ and the PpeI promoter) caused recombination, resulting in the deletion of FleQ. 

Figure 8: The three fragments with sticky ends were used in Golden Gate assembly with SapI. Created with BioRender.

Figure 9: Ligation of pSB1A3 with only Promoter as an insert. Created with BioRender.

A third design was made for a Golden Gate Assembly with BsaI. BsaI cleavage results in sticky ends of 4 bases, instead of the 3 bases with SapI. This increases the amount of possible combinations from 64 to 256. Because the pSB1A3 vector contains a BsaI restriction site in the ampicillin resistance marker, the construct was designed with the pSB1K3 vector. In order to remove the repeating sequence, the Pcl promoter in front of FleQ was replaced by the high-strength BBa_J23100 from the Anderson constitutive promoter family. As a high concentration of FleQ is required to detect above-native levels of cyclic-di-GMP, a strong promoter is required. Additionally, the BBa_J23100 has been well characterized in E. coli TG1 (http://parts.igem.org/Part:BBa_J23100). This new promoter sequence was placed on the primer overhangs and the fragment was amplified behind the Pcl promoter in order to remove it. All the fragments required for the assembly were successfully obtained by Q5 PCR, as confirmed by 0.9% agarose gel. Colonies obtained after transformation with the assembly reaction mix were screened by colony PCR and restriction digest (NotI) but no right fragment sizes were identified.

Figure 10: Binding site of FleQ_Pro_BsaI_Fwd primer on synthetic DNA fragment. Image from Benchling.

Figure 11: The sticky ends of the fragments created by BsaI digest for Golden Gate Assembly. Additional effort was put in to reduce the compatibility of the sticky ends between undesired combinations.

To create Construct 4, our finalized biosensor module, our reporter module (described at Construct 1) and our sensor module (described at Construct 3) needed to be combined. As described, we were unable to obtain Construct 3 due to recombination problems resulting from the use of various repeating sequences. If we had been successful we would have been able to put a truncated version of Construct 3 (lacking the RFP element) before Construct 1 (GFP + T) and subsequently test the construct. The theoretical part is denoted in the part registry as BBa_K4730021. 

Figure 13: Plasmid map of designed Construct 4 (Part Code: BBa_K4720021). Part composition: BBa_J23100 + BBa_K4720000 (FleQ) BBa_B0015 (TT) + BBa_R0051 (Pcl) + BBa_K4242001 (Ppel) + BBa_J34801 (RBS) + BBa_E0040 (GFP) + BBa_B1006 (Terminator). Benchling link

Ultimately, our biosensor should contain the full length of our sensor and reporter modules. Above, our final product is described. However, the RFP under the control of a constitutive promoter is still missing. Additionally, in the current design of our final construct, there is still the risk of recombination, as seen throughout various of our experiments. Currently, FleQ, the GFP and the RFP fragments are all under, partial, control of the constitutive promoter PcI. Besides this, there is also the issue of the double use of terminator BBa_B0015.

As an alternative to the terminator BBa_B0015, terminator BBa_1002 has been selected. To replace the constitutive promoters, both BBa_J23111 and BBa_J23118 have been identified. These promoters are of interest due to their similarities in promoter strength (see Anderson collection). Note that there is homology between the three Anderson promoters used, but the promoters are shorter than the PcI promoter and there are small variations. This will need to be confirmed with additional experiments, but the thought is that these changes are sufficient to overcome the problem of recombination. 

Construct 4B: Biosensor control modules (Construct 1+3B, control for FleQ)

In order to test if the Pcl + Ppel tandem promoter would work in our E. coli strain TG1 we constructed the same plasmid as above, but without the FleQ gene. This plasmid was made by the standard assembly (RFC[10]) of BBa_K4720018 in front of pSB1C3_BBa_K4720012. This construct was successfully obtained but needs purification. The sequencing results show two different sequences present behind the Pcl + Ppel promoter, with one being right. In order to isolate the right construct, a dilution streak could be made with subsequent screening of the isolated colonies for the right construct. This was however not done due to the time constraints.

Ultimately our project would have resulted in the creation of three distinct plasmids:

1. The first plasmid would have been our combined sensor and reporter modules, a cyclic-di-GMP dependent GFP with a constitutively expressed RFP. This plasmid would, besides continuously expressing RFP, express GFP in the presence of the biofilm-related signaling molecule. If cyclic-di-GMP is sensed, it will bind to the repressor FleQ, changing its affinity for the Ppel binding site and activating the promoter by no longer blocking RNA polymerase from binding (BBa_K4720106).
2. In our biofilm sensor, our E. coli biosensor will be immobilized and, thus, might be producing cyclic-di-GMP itself. Additionally, there is a native concentration of cyclic-di-GMP present. A problem was also noted, and tackled, by the CUG China team in 2022. Our second plasmid, just like CUG’s design, contains a cyclic-di-GMP hydrolase to reduce the natively present cyclic-di-GMP to ensure that our biosensor only activates upon binding of cyclic-di-GMP that originates from the biofilm (BBa_K4720022).
3. Our third and final plasmid contains our treatment or response module. After many considerations on how to use the M13 phage, we settled on the PIII phage display of the biofilm-degrading enzyme Dispersin B. The PIII and Dispersin B are under the same cyclic-di-GMP dependent control as the GFP. Thus, if a biofilm is present, not only the reporter will be activated, but also our Dispersin B will be expressed. To be able to generate a functional M13 phage displaying the Dispersin B we will also need to infect our biosensor with a helper phage, containing all genes necessary to generate a fully functional M13 phage. 

1. Kong, W. C. (2015, August 28). Part: BBa_K1659210. Registry of Standard Biological Parts.
2. Breslawec, A. P., Wang, S., Monahan, K. N., Barry, L. L., & Poulin, M. B. (2023). The endoglycosidase activity of Dispersin B is mediated through electrostatic interactions with cationic poly-β-(1→6)-N-acetylglucosamine. The FEBS journal, 290(4), 1049–1059.
3. Wang, R., Li, H. D., Cao, Y., Wang, Z. Y., Yang, T., & Wang, J. H. (2023). M13 phage: a versatile building block for a highly specific analysis platform. Analytical and bioanalytical chemistry, 415(18), 3927–3944.
4. Reinders, A., Sellner, B., Fadel, F., van Berkum, M., Kaczmarczyk, A., Ozaki, S., Rueher, J., Manfredi, P., Sangermani, M., Harms, A., Perez, C., Schirmer, T., & Jenal, U. (2021). Digital control of c-di-GMP in E. coli balances population-wide developmental transitions and phage sensitivity [Preprint]. Microbiology.
5. New England Biolabs. (n.d.). M13MP18 RF i DNA | NEB.
6. New England Biolabs. (n.d.-a). M13KO7 Helper Phage | NEB.
7. Wang, R., Li, H., Cao, Y., Wang, Z., Yang, T., & Wang, J. (2023b). M13 phage: a versatile building block for a highly specific analysis platform. Analytical and Bioanalytical Chemistry, 415(18), 3927–3944.
8. Libretexts. (2021). 4.2: M13 Phage. Biology LibreTexts.