Main engineering goal - The main engineering goal of this project is to develop a new synthetic biology tool that could be applied to the detection of different nanoplastics - nanoparticles of polystyrene (PS), polypropylene (PP) and polyethylene (PE). Our unique tool will be based on antimicrobial peptides that have different affinities toward various plastic materials. Moreover, we aim to describe the ways of how such a system could be used in future real-world scenarios.

Main results - Our developed nanoplastics detection tool successfully provides a plastic-concentration dependent fluorescent signal for PS and PP.

nP - nanoplastic, e.g. nPS, nPP, nPE.

CBD-PBD - a fusion peptide that has both cellulose binding domain and Plastic Binding Domain, i.e. CBD-TA2 and CBD-LCI

GFP-PBD - a fusion peptide that has both plastic binding domain and green fluorescent protein, i.e. GFP-TA2 and GFP-LCI

CSD-PBD - E. coli bacteria that are displaying plastic binding peptides on their surface, i.e. CSD-TA2 and CSD-LCI.

In order for our detection system to be practical to use in real world scenarios it is important that the main component of the assay - plastic binding peptides - can be produced in a simple and efficient manner. For this reason we have screened different peptide production parameters: strain, growth time, growth temperature, IPTG concentration, pH of lysis buffer. In order to determine the best way to produce each peptide. Table 1 demonstrates the best conditions for each case.

After the best growth and lysis conditions were found, peptides were purified using nickel affinity chromatography columns. SDS-PAGE of cell free extracts and purified peptides is shown in Fig. 1.

Fig. 1 Cell free extracts and purified peptides. L - protein ladder (PageRuler™ Plus Prestained Protein Ladder #26619), C - cell free extract of no-plasmid control, S - cell free extract of, P - purified and peptide.

However, even when the best conditions were used, peptides had shorter bands besides the main band, indicating possible degradation. We assumed this could be due to proteases that could not be inhibited by the protease inhibitor cocktails we used. We have checked this by incubating peptide solution of purified CBD-TA2 with the cell free extract of untransformed E. coli BL21 DE3 bacteria. The intensities of the two bands did not seem to change even after 1h of incubation (Fig. 2). Thus, there had to be other reasons for the occurence of shorter bands, besides degradation due to proteases.

Fig. 2 CBD-TA2 degradation assay. Cell free extract of E. coliBL21 DE3 (no-plasmid control) was added to a sample of purified CBD-TA2 and the mixture was incubated for 10, 20, 30, 40, 50 and 60 min. at 37℃. L - protein ladder (PageRuler™ Plus Prestained Protein Ladder #26619), K - purified CBD-TA2 peptide, incubated for 60 min. at 37℃.

To find out whether CBD-PBD peptides bind to plastics, CFEs were incubated with macro-sized pieces of PS, PP and PE. After washing the unbound fraction, bound peptides were eluted. A faint band in the B line could be seen for all peptide-plastic combinations, confirming that both CBD-TA2 and CBD-LCI peptides bind to PS, PP and PE plastics (Fig. 3).

Fig. 3 CBD-TA2/LCI binding to macroplastics. A, peptide binding to PS; B, binding to PP; C, binding to PE. L - protein ladder (PageRuler™ Plus Prestained Protein Ladder #26619), S - cell free extract of CBD-TA2 or CBD-LCI, U - unbound fraction, W - washed fraction, B - bound fraction of peptides.

Additionally, we had noticed that CBD-TA2/LCI peptides may be binding to the hydrophilic PVDF membrane that we use to filter elutions from cellulose and plastic particles (Fig. 4).

Fig. 4 PS bound fraction of CBD-TA2 and CBD-LCI taken before (-) and after (+) centrifugation through hydrophilic PVDF membrane. L - protein ladder (PageRuler™ Plus Prestained Protein Ladder #26619).

GFP-TA2/LCI binding to PS, PP and PE was tested using the same method. However, Western blotting (WB) was needed to see the faint bands due to low concentration of peptides in the CFE (Fig. 5). The same CFEs, whose profiles are shown in Fig. 5, were used. The bands marked with green arrows were confirmed to be the CBD-TA2 and CBD-LCI peptides, as their intensity in WB compared to the whole protein profile in the same line of SDS-PAGE increased substantially. Bands seen in washed (W) and/or bound (B) fractions were considered to indicate protein binding. This way GFP-TA2 was confirmed to bind to PS, PP and PE, and GFP-LCI - to PS and PP. Nevertheless, the data has to be assessed carefully due to the antibody binding to nonspecific bands in the no-plasmid control and the overall low intensities of protein bands. Moreover, bands of incorrect size could be seen in the washed and bound fraction (marked with blue and gray arrows), with the lower size bands indicating possible peptide degradation.

Fig. 5 GFP-TA2/LCI binding to plastics. A, peptide binding to PS; B, binding to PP and PE. L - protein ladder (PageRuler™ Plus Prestained Protein Ladder #26619), S - cell free extract of no-plasmid control, U - unbound fraction, W - washed fraction, B - bound fraction of peptides. Green arrow - bands of the right size (~35.0 kDa for GFP-TA2 and ~35.5 kDa for GFP-LCI) seen after washing/elution. Gray and blue arrows - incorrect size bands seen after washing/elution.

To be able to immobilize CBD-TA2 and CBD-LCI peptides on cellulose, we first checked whether it binds to it. Cell free extracts that contain the peptides were incubated with Whatman paper which was cut into small pieces. After washing the unbound fraction, bound peptides were eluted. A faint band can be seen in the bound fraction for both CBD-TA2 and CBD-LCI (Fig. 6). Furthermore, the amount of peptide that was removed during the washing step was comparably lower than in the bound fraction. This showed that both peptides could successfully be immobilized on Whatman paper.

Fig. 6 CBD-TA2/LCI binding to Whatman cellulose paper. L - protein ladder (PageRuler™ Plus Prestained Protein Ladder #26619), S - cell free extract of CBD-TA2 or CBD-LCI, U - unbound fraction, W - washed fraction, B - bound fraction of peptides.

In order to test whether our detection system could potentially work with naturally formed nPs, we decided to make nPs in a more natural manner - mechanical fragmentation. PS, PP and PE nanoplastics were made mechanically by grinding. First, dry pallets were ground using a coffee grinder, then suspended in ethanol and ground using a planetary ball mill.

The average sizes of particles, measured using atomic force microscopy, were from 95.5 nm to 341,5 nm in length and from 21.2 nm to 37.8 nm in height (sizes for each type of nPs shown in Table 2). Big particles of PP and PE with a diameter of 493 nm and 550 nm, respectively, could be seen and indicate possible aggregation (Fig. 7). This could have had a negative impact on the precision of the size measurements.

Fig. 7 Atomic force microscopy of PS, PP and PE plastics after mechanical fragmentation.

In order to use the nPS in detection experiments, 0,05 % w/v aqueous dispersions were made in 5 mM Tris-HCl pH 7.4, 1 mM NaCl, 0.05 % v/v LAS surfactant. These dispersions were left to set for a few days to form layers with different particle stability. Then, the stable middle layer was separated for use in detection experiments (Fig. 8).

Fig. 8 Aqueous colloidal dispersions of mechanically fragmented nPS, nPP and nPE made for use in detection experiments.

The size distributions of plastic particles in aqueous dispersions were measured by dynamic light scattering (DLS; Fig. 9).The data show that all three dispersions of plastic particles (PS, PP and PE) had a size peak in the nanometer range.

Fig. 9 DLS measurements of hydrodynamic radius of PS, PP and PE particles in aqueous dispersions. Error bars show standard deviation of three measurements.

Testing of the detection system was done using glass coated microplates. At the bottom of the well a round piece of Whatman paper (grade 1; with a diameter of 5.5 ± 0.5 mm) was placed and CBD-PBD peptide was immobilized on cellulose. Unbound cellulose was washed 3 times with 50 mM Tris-HCl, pH 7.4. Then, an aqueous colloidal dispersion of nPs was placed inside the well and incubated to bind to the immobilized CBD-PBD. After the well was washed 3 times, either GFP-PBD peptides or CSD-PBD bacteria were added for a fluorescence response. Unbound peptides or bacteria were washed 3 times if not indicated differently, and the signal was read with a microplate reader. Please refer to the protocol book for more detailed protocols.

Three peptide and CSD bacteria combinations were tested:

• CBD-TA2 and GFP-TA2
• CBD-LCI and GFP-LCI
• CBD-TA2 and CSD-TA2

To make sure that the background signal is as low as possible, we have tested whether the cellulose that we use for CBD-PBD immobilization has any autofluorescence. First, background fluorescence of a microplate well was measured. Then, a piece of Whatman paper (of the same size, as used in detection experiments) was placed in a glass coated microplate. Fluorescence was measured before and after washing with mQ water. As seen in Fig. 10, cellulose does emit a relatively high fluorescence of 520 nm when excited at 485 nm, but after the washing step, its intensity reduced almost to the background.

Fig. 10 Autofluorescence of Whatman filter paper (grade 1, with a diameter of 5.5 ± 0.5), measured in a well of a glass coated microplate. Error bars show standard deviation of fluorescence from three measured wells.

We have noticed that when a smaller amount of CBD-TA2 is added on cellulose for immobilization, higher fluorescence is seen in the no-nPS control (Fig. 11). In this experiment, either 0.4 nmol, 2 nmol or no CBD-TA2 was added on a piece of cellulose paper in a microplate well. There, commercially obtained 100 nm ± 10 nm size negatively charged nPS particles was used. The result indicated that GFP-TA2 also strongly binds to Whatman filter paper. Therefore, saturation of Whatman paper with CBD-PBD or other type of blocking was necessary.

Fig. 11 Dependence of fluorescence intensity on the amount of CBD-TA2. Each condition was tested once.

To find the right amount of CBD-TA2/-LCI for immobilization on cellulose, that would lower the background signal significantly, cellulose saturation experiments were done. Results have showed that when the amount of added CBD-TA2 or CBD-LCI is higher (e.g. around 3 nmol, or 80 µg, for both peptides) the quantity of bound protein is growing slower, i.e. the curve is slightly bent (Fig. 12). Although the maximum amount of added peptides to cause full saturation of a cellulose paper with a 5.5 mm ± 0.5 mm diameter was not shown by this experiment. For detection experiments in microplates we chose to add 3.4 nmol (or 90 µg) of both CBD-TA2 and CBD-LCI peptides.

Fig. 12 Whatman paper saturation with CBD-TA2 (A) and CBD-LCI (B). Each amount of added CBD-TA2 and CBD-LCI was tested in triplicate. The curves were fitted by applying local polynomial regression (R function loess).

The first detection test was done using CFEs that contained CBD-TA2 and GFP-TA2 peptides. Thus, in this test, the exact amount of added CBD-TA2 and GFP-TA2 was not known. As a nanoplastics sample, different concentrations of commercially obtained 100 nm size negatively charged nPS was used: 0.5%, 0.1%, 0.2% of solids, and a buffer solution without nPS (no-nPS control). 100 µl of each component of the detection system was added for each step of the procedure (i.e. immobilization of CBD-TA2, washing, addition of nPS and addition of GFP-TA2). Results showed that fluorescence intensity depends on the concentration of nPS (Fig. 13).

Fig. 13 Concentration-dependent fluorescence intensity when CBD-TA2 and GFP-TA2 peptides were used for the detection of commercially obtained nPS. Error bars show standard deviation of two technical repeats.

We then tested the nPS that were made by mechanical fragmentation. This time both CFEs (data not shown) and purified peptides were used for nPS detection. When CFEs were tested, higher fluorescence intensity in the no-nPS control was observed compared to samples with nPS. We speculated that impurities in the CFEs could have caused nonspecific interactions that led to a higher signal in the no-nPS sample.

Therefore, we decided to do the detection test using purified peptides (Fig. 14). In this experiment, 90 µg of CBD-TA2 per well was used for immobilization and 10 µg of GFP-TA2 was added to give the fluorescent signal. The nPS dispersion was diluted 10⁴, 10⁵, 5*10⁵, 7,5*10⁵, 10⁶ and 10⁹ times. Results showed that similarly to the when CFEs were used, the no-nPS control emitted higher signal compared to the wells in which nPS was added, except for the 7,5*10⁵ dilution. This meant that at some concentrations mechanically ground nPS could be detected. We hypothesize that the 7,5*10⁵ dilution of nPS could be the maximum concentration that can still be detected and that higher concentrations could initiate nPS aggregation, which in turn would cause inaccuracy of measurements.

Fig. 14 Detection of mechanically fragmented nPS using purified CBD-TA2 and GFP-TA2. Error bars show standard deviation for measurements of two technical repeats.

Moreover, the difference in the results, that were generated using the commercial and our own nPS, shows that charge and/or size distribution of nPs may be important for the sensitivity of detection.

The second system that was tested - detection of mechanically fragmented nPP using CBD-LCI and GFP-LCI peptides. Purified peptides were used in this experiment (90 µg of CBD-LCI and 10 µg of GFP-LCI per well). The nPP dispersion was diluted 10, 20, 50, 10², 10³, 10⁴, and 10⁶ times (Fig. 15, A). The no-nPP control showed a higher signal compared to all nPP dilutions except for the sample that was diluted 10⁶ times. This result supports the data that was acquired during experiments with mechanically fragmented nPS and shows that high concentrations of nPPs can not be detected as well.

A second test was done with nPP of higher dilutions (10⁴, 10⁵, 5*10⁵, 10⁶ times diluted)(Fig. 15, B). This time low signal was emitted from the no-nPP wells, and a peak at the 10⁵ dilution could be seen, again, supporting the hypothesis that our detection system has a maximum nPs concentration limit.

Fig 15. Two separate experiments showed concentration-dependent signal for nPP. The signal peak was at 10⁹ dilution in the first eksperiment (A), and at 5*10⁵ in the second (B). Each sample was tested twice in both experiments.

After showing that phusion peptides can be used for nPs detection, we decided to test our system using E. coli bacteria that display plastic binding peptides on their surface (CSD-PBD) as an alternative to GFP-PBD. For CSD, we used a mutated version of Pseudomonas aeruginosa estA and fused it to TA2 and LCI peptides together with an e-epitope for antibody recognition (see parts BBa_K4380020 and BBa_K4380019, respectively). These cells were therefore called CSD-TA2 and CSD-LCI. Additionally, the cells were cotransformed with a pACYC184 plasmid that constitutively expressed either eGFP, or mScarlet proteins under the tetracycline promoter. Control cells (CSD-C) were made by fusing estA only with an e-epitope.

Proof that E. coli cells are displaying peptides on their surface was done by incubating bacteria with anti-e antibodies that would attach to the e-epitope which is incorporated in the sequence of the displayed peptide. The FITC molecules that were attached to anti-e antibodies allowed to see cells using fluorescence microscope. Fluorescence microscopy in Fig. 16 shows that when CSD-C+mScarlet and CSD-LSI+mScarlet cells were induced with IPTG, some cells appeared both green and red when excited at 485 nm and 565 nm, respectively. On the other hand, when uninduced cells were excited, the cells appeared to emit only red fluorescence. For the CSD-C+eGFP and CSD-TA2+eGFP bacteria, another method had to be used, as green fluorescence was already produced by eGFP. Nevertheless, using this method we proved that CSD-C+mScarlet and CSD-LCI+mScarlet bacteria are displaying peptides on their surface.

Fig. 16 Fluorescence microscopy of CSD bacteria. Induced and uninduced CSD-C+eGFP, CSD-TA2+eGFP, CSD-C+mScarlet, CSD-LCI+mScarlet cells and uninduced no-plasmid control cells were examined under microscope.

A more quantitative test was done in order to show that CSD-C+eGFP and CSD-LCI+eGFP bacteria also display the peptides. Culture samples were taken from both induced and uninduced bacteria and their OD was equalized. Then equal amounts of the samples were washed with PBS and incubated with anti-e antibodies. Fluorescence of eGFP or of both eGFP and FITC was measured (Fig. 17).

Fig. 17 Fluorescence intensity measurements of induced and uninduced CSD-C+eGFP, CSD-TA2+eGFP, CSD-C+mScarlet, CSD-LCI+mScarlet cells and uninduced no-plasmid control cells that were or were not incubated with anti-e antibody. Error bars show standard deviation of triplicates.

Results showed that in all cases, except for the induced sample of CDS-C+eGFP, fluorescence increase was higher compared to the no-plasmid control when anti-e antibody was added (Fig. 18). This data proved that CSD-TA2+eGFP are displaying peptides. However, uninduced cells of both CSD-C+eGFP and CSD-TA2+eGFP had a bigger difference of fluorescence before and after addition of antibodies which indicated that uninduced cells were displaying comparably more peptides than the induced cells.

Fig. 18 Difference of fluorescence intensity of cells that were or were not incubated with anti-e antibodies. Induced and uninduced CSD-C+eGFP, CSD-TA2+eGFP, CSD-C+mScarlet, CSD-LCI+mScarlet cells and uninduced no-plasmid control cells were tested. Error bars show standard deviation of triplicates.

To find out whether higher expression of peptides caused the bigger increase in fluorescence in case of CSD-C+eGFP and CSD-TA2+eGFP cells, Western blot using the anti-e antibody was done (Fig. 19). SDS samples were made from cultures with equalized optical density. Results show that after induction with IPTG, expression of all CSD constructs was increased, which did not support the hypothesis that the higher amount of peptides could be displayed due to higher peptide expression.

Fig. 19 Western blot (A) and SDS-PAGE (B) of estA phusion proteins. Samples were taken from cultures that were or were not induced for protein expression. L - protein ladder (PageRuler™ Plus Prestained Protein Ladder #26619).

To be able to use CSD-PBD cells for detection experiments, we first had to test whether cellulose emits any fluorescence at 594 nm when excited at 565 nm. Same procedure was done for measuring autofluorescence at 485 nm. As seen in Fig. 20, cellulose does emit a fluorescence at 594 nm, but after washing with mQ water, its intensity reduced almost to the background.

Fig. 20 Autofluorescence of Whatman filter paper (grade 1, with a diameter of 5.5 ± 0.5), measured in a well of a glass coated microplate. Error bars show standard deviation of fluorescence from three measured wells.

As an alternative to GFP-PBD for the generation of fluorescence, bacteria that display peptides on their surface were used (i.e. CSD-LCI bacteria). In this detection test, only 18 µg of CBD-TA2 per well was used for immobilization and 200 µl of CSD-LCI bacterial culture was used to produce fluorescent signal. The nPP dispersion was diluted 10, 50 and 250 times. The 10 and 50 times diluted nPP gave a higher signal compared to noPP control and other control samples (only CSD-LCI, noCBD-LCI and no-nPP, no-CBD-LCI, no-nPP)(Fig. 21). This information gives rise to speculation that CSD bacteria have less nonspecific interactions than GFP-LCI peptide, or some other properties enable their use for the detection of high concentrations of nPs.

Fig. 21. Detection of mechanically fragmented nPP using purified CSD-LCI bacteria. Each sample was tested once.

epPCR - error-prone PCR

8-Oxo-dGTP - (8-Oxo-2'deoxyguanosine-5'-triphosphate)

dPTP - (6H,8H-3,4-Dihydro-pyrimido(4,5-c)(1,2)oxazin-7-one-8-β-D-2'-deoxy-ribofuranoside-5'-triphosphate)

Although using known peptides for a detection system already is a huge challenge in itself, we wanted to characterize novel peptides and create huge peptide libraries, which could be used for specific plastic binding in our detection system. Therefore, we decided to do a peptide evolution, specifically designed for plastic-binding peptides. Although robust evolution methods create huge libraries, in order to specifically tailor one's function to bind a specific substrate stronger, a high mutation rate is required.

epPCR technique was performed using two different strategies (see Design or Engineering pages for more information):

• 1) epPCR using modified oligonucleotides 8-Oxo-dGTP and dPTP.
• 2) epPCR utilizing low fidelity of Taq polymerase.

epPCR of the specific peptide sequence was carried out in the presence of 0,2mM of dNTP, 4uM of dPTP and 0,1 mM 8-oxo dGTP. 35 cycles of amplification were performed with 0,5 uM of each epPCR primer (Parts: BBa_K4380006, BBa_K4380007) during the 1st PCR amplification cycle.

Mutations were generated and the megaprimer method was utilized as a way to clone our mutated sequences to plasmid vectors. After the 2nd round, 500 ng of Megaprimer generated in the 1st round was used as a primer for the second round of PCR onto bacterial plasmid vector. Bacterial colonies were isolated, plasmid were purified and 6 of the plasmids were sent for sequencing.

epPCR of the specific peptide sequence was carried out by utilizing commercial kit (PickMutant™ Error-Prone PCR Kit), which is based on the protocol of Cadwell and Joyce (1992). The experiment was carried out according to the manufacturer's recommendation, but double the amount of MnCl2 was used.

epPCR of the specific peptide sequence was carried out by taking advantage of the inherently low fidelity of Taq DNA polymerase, which may be further decreased by the addition of Mn²⁺, increasing the Mg²⁺ concentration, and using unequal dNTP concentrations. The method, utilized here was composed of 0,5 mM MnCl2, 7 mM MgCl2, dCTP and dTTP 1mM each, dATP and dGTP at 0,2 mM each.

Mutations were generated and the megaprimer method was utilized as a way to clone our mutated sequences to plasmid vectors. After the 2nd round, 500 ng of megaprimers (fragments generated during the 1st round of PCR) were used as a primer for the second round of PCR onto bacterial plasmid vector. After successful cloning and transformation, colonies were formed. Bacterial colonies were isolated, plasmids were purified and 6 of the plasmids were sent for sequencing.

The sequencing results were:

Although our megraprimers were successfully cloned into bacterial vectors, the technique is not suitable for this protocol. During the second round of PCR on a plasmid vector, the vector itself gets mutated because, as we believe, there are not sufficient enough methods to purify a highly mutated megaprimer, and only colonies with unmutated antibiotic resistance gene survive.

In the end, we have decided to skip the directed evolution after the first few iterations and prioritize other aspects of the project, as the other peptides already used in the system provided a satisfactory binding. In the future, to fully optimize the detection system for potential commercial or research applications we would suggest to reiterate a directed evolution approach and possibly gain even more potent binders.