Team:UCL/Engineering

Team:UCL/Page Name - 2020.igem.org

Overview

The engineering design of PETZAP started in April 2020 and involved an iterative design as demonstrated in the following figure. This design cycle integrated our team’s work in human practices, synthetic biology, and modeling.

On this page, we are presenting how the design cycle contributed to the three key engineering aspects of our project: Construct Design, Coculture Design, and Microbial Desalination Cell (MDC) Configuration Design.

Part 1: Construct Design

Overall Aim

Our Synthetic Biology group was dedicated to designing constructs that would allow our bacteria to acquire the desired functionalities to fully degrade PET and feed an exoelectrogenic bacterial strain, Shewanella oneidensis, in the Microbial Desalination Cell. This is how it would work: Escherichia coli will degrade PET to TPA while Pseudomonas putida will fully metabolise TPA and produce lactate for S. oneidensis. The following demonstrates the metabolic pathway in the two plastic degrading species which our construct design work focuses on.

Figure 1. Pathway schematic representation of a model of enzymatic plastic degradation.
The engineered E. coli secretes PET degrading enzymes, PETase and MHETase, to degrade polyethylene terephthalate (PET) into mono-terephthalic acid (MHET) and ethylene glycol (EG) [2][3]. EG enters E. coli via transporters to be further metabolized in the native secondary pathway expressed an operon in E. coli. While glycerate (GL) enters glycolysis and gets converted into pyruvate (Pyr) to support E. coli’s biomass growth, TPA enters P. putida via TPA transporter encoded by genes from Comamonas sp. Strain E6 [19]. Our engineered P. putida expresses transgenic genes to degrade TPA into 1,2-dihydroxy-3,5-cyclohexadiene-1,4-dicarboxylate (DCD), next into 3,4-dihydroxybenzoate (PCA) [17]. PCA enters P. putida’ s endogenous β-ketoadipate pathway to be metabolized into succinyl-CoA (SucCoA), which enters tricarboxylic acid cycle (TCA) of P. putida, generating lactate for exoelectrogen S. oneidensis.

For the complete metabolic pathway including S. oneidensis, please visit our Model page.

Objective 1: Expressing PET degradation pathway in E. coli

Our first objective of construct design was to degrade PET bottles collected from the polluted ocean water using E. coli BL21 (DE3) expressing the cooperative enzymes, PETase and MHETase.

PETase and MHETase have been successfully expressed in E. coli [1] [2]; however, rather than secreting these two enzymes individually, we decided to create a fusion protein to possibly increase catalytic turnover. This strategy has been commonly used in biocatalysis since it ensures that the product of PETase readily becomes the substrate for MHETase.

We first decided to do some structural modelling prior to designing the constructs, since we had doubts about the gene order for the fusion protein. For this, we made use of the structural prediction software, Phyre2 [3] as well as Pymol [4]. In this way, we would identify possible clashes and infer which fusion protein may be more stable.

PETase-MHETase fusion protein Phyre2 structural prediction

MHETase-PETase fusion protein Phyre2 structural prediction

Figure 1. Structural modelling of PETase-MHETase and MHETase-PETase.
(A) Structure of PETase-MHETase. The hydrophilic loops found between MHETase (purple) and PETase (cyan) correspond to the disordered N-terminus of this enzyme. These are predicted to fold close to the hydrophilic flexible Gly-Ser linker (green), thereby positioning PETase on top of the active site of MHETase. The dynamic behaviour of the flexible linker and the disordered N-terminus may cause aggregation for this configuration as the fusion protein is synthesised by the ribosome. Similarly, the MHETase catalytic triad is at the centre of the structure and the MHET polymer is known to go through the middle pore of MHETase to be degraded. Hence, it is possible that the positioning of PETase might block the active site entrance, thereby leading to reduced activity of the overall fusion protein. (B) Structure of the MHETase-PETase. The two enzymes are predicted to be positioned further apart, since now the disordered N-terminus of the MHETase is not in close proximity to the linker. The longer N-terminus of PETase when fused to the linker ensures sufficient distance between both proteins. It is also noticeable that catalytic centre of MHETase is accessible in this configuration. Therefore, the MHETase-PETase construct is predicted to perform better.

Later, our human practices group contacted the McGeehan group [5], which was about to publish a paper discussing different PETase-MHETase fusion protein configurations. Their results showed that expression of PETase-MHETase was not possible due to protein aggregation, however, they did manage to solubly express a MHETase-PETase fusion which was catalytically active in E. coli. Our team decided to further verify his hypothesis by designing two constructs with different gene order for expression in E. coli.

In the native organism, Ideonella sakaiensis, PETase and MHETase are secreted extracellularly [1] through the Sec-dependent pathway [6]. We identified the native N-terminal signal peptide sequence for PETase and MHETase, through analysis with the Signal Peptide prediction software, Signal P-5.0 [7]. and predicted their respective cleavage sites (Fig. 1 and 2).

PETase results

Figure 1. PETase prediction of signal peptide cleavage site using SignalP-5.0.

From the sequence analysis, the most likely cleavage site is between amino acids 27 and 28, with the Sec/SPI being the most likely secretion pathway in a Gram-negative bacterium like E. coli with 77% likelihood.

MHETase results

Figure 2. MHETase prediction of signal peptide cleavage site using SignalP-5.0.

From the sequence analysis, the most likely cleavage site is between amino acids 17 and 18, with the Sec/SPII being the most likely secretion pathway in a Gram-negative bacterium like E. coli with 99% likelihood.

After identifying the length of the native signal peptide sequence, these were replaced to allow efficient secretion by our chosen chassis, E. coli.

After contacting Prof. John Ward from the Biochemical Engineering department at UCL and conducting literature research, we found evidence that both PETase and MHETase were secreted independently in E. coli through the Sec-dependent pathway (Fig. 3) when using the N-terminal signal peptide, LamB [2]. Therefore, we attempted the same strategy for the extracellular secretion of the fusion protein, by choosing LamB as the N-terminal signal peptide.

Figure 3. Sec-dependent pathway protein secretion.

The Sec-dependent pathway was the first secretion pathway to be discovered in bacteria in charge of directing proteins to the periplasm and the cell membrane by translocating unfolded nascent chains across the inner membrane [6]. The reducing environment of the periplasm allows formation of disulphide bonds which allow proteins to fold. Furthermore, a variety of chaperones ensure correct protein folding in the periplasm before proteins can be secreted extracellularly through either the Bam, Omp or FimD complexes. This pathway requires an N-terminal signal sequence that allows translocation before it is cleaved.

Next, we had to choose a linker for our fusion protein. After considering both a rigid and a flexible linker, we made the choice of using a flexible Gly-Ser linker [8] to allow the two enzymes to retain activity even after being fused together. Similarly, we thought that a flexible linker would interfere less during the folding process in the periplasm.

We wanted to test if the gene order of the enzymes on the fusion protein affected extracellular secretion, expression or activity [5]. This led to the design of constructs 1-4 (Fig. 4). Constructs 1 and 2 were the fusion protein with different gene order for PETase and MHETase, while constructs 3 and 4 served as positive controls by expressing the free enzymes. This set up allowed us to compare the activity of the fusion protein against the free enzymes as a control and determine if there was a fold increase in activity. The empty plasmid (T1_FB) was used as the negative control.

All 4 constructs were codon optimized for expression in E. coli K12 using the Benchling codon optimization tool, which makes use of the harmonization algorithm to account for codon usage burden as opposed to other algorithms that always use the most frequent codon. In this way, we ensured that our cells would not suffer from amino acid depletion.

All the constructs contained LamB as the N terminal signal peptide and a C terminal 6xHis tag. The 6xHis tag would facilitate the purification step [9] for the Exeter team and allow them to perform a clean and reliable enzyme activity assay (pNPA assay) [10] on all 4 constructs. The pNPA assay would indicate which gene order has the highest PETase activity.

Figure 4. Transforming the 4 constructs into E. coli BL21 cells.

All 4 constructs were transformed into E. coli cells. Colonies were identified thanks to the antibiotic resistance marker, chloramphenicol, present in all 4 constructs. The N-terminal LamB signal peptide is shown in pink, PETase in blue, MHETase in red and the flexible Gly-Ser linker in green. Constructs 1was the PETase-MHETase fusion, while construct 2 was the MHETase-PETase. Constructs 3 and 4 were the positive controls for the soluble PETase and MHETase, respectively.

We created 4 constructs in BioBrick [11] format to contribute to the registry. Each of the 4 constructs was:

Flanked by the corresponding prefix and suffix of this standard
Codon optimized for expression in E. coli
Checked so no forbidden sites were present within the BioBrick
Synthesised by GenScript and cloned by GenScript into the plasmid T1_FB.

Figure 1. The 4 constructs in Biobrick standard.

All 4 constructs are shown in SBOL standard and have been flanked by the corresponding prefix (EcoRI and XbaI) and suffix (SpeI and PstI) of the Biobrick standard [3], in order to facilitate sharing of the constructs among the scientific community. The common features to all constructs were a constitutive T7 promoter, a strong RBS, the N-terminal signal peptide LamB, a T7 terminator and a 6x His Tag for purification. Constructs 1 and 2 are the two versions of the fusion protein that were tested with the para-Nitrophenyl acetate (pNPA) assay to determine if changing the order of PETase and MHETase had an effect on enzyme activity. Construct 3 and 4 are the positive controls for PETase and MHETase, respectively.

Although our team has no access to wet lab this year due to the global pandemic, we are proposing the following protocols for our partner, Exeter, to test the four E. coli constructs in their lab.

Section 1. Cloning of PETase, MHETase into plasmid T1_FB

Clon the DNA sequence of PETase, MHETase and their fusion into the plasmid T1_FB using SpeI and NdeI as the restriction sites.

Figure 1. Plasmid map of T1_FB on Benchling with restriction sites SpeI shown in orange and NdeI shown in green.

Figure 2. Four proposed parts.
Construct 1 (BBa_K3601000). PETase-MHETase fusion protein with a GS flexible linker jointing the C-terminus of PETase and N-terminus of MHETase, and a LamB secretion peptide on the N-terminus of PETase. Construct 2 (BBa_K3601001). MHETase-PETase fusion protein with a GS flexible linker jointing the C-terminus of MHETase and N-terminus of PETase, and a LamB secretion peptide on the N-terminus of MHETase. Construct 3 (BBa_K3601002). LamB secretion peptide with PETase alone. Construct 4 (BBa_K3601003). LamB secretion peptide with MHETase alone.

Section 2. Transformation and E. coli cell growth protocols

We proposed to use E. coli BL21 (DE3) as our expression strain for each of the 4 constructs. During the initial process of cell culture, we planned to supply glucose to ensure we would have sufficient number of cells before switching to give PET film as the only food source while for cells transformed with construct 4 (MHETase alone), we would supply MHET rather than PET films after initial culture on glucose.

Section 3. Protein purification protocols

After cell growth, each of the samples of the 4 different culture media would be taken, and secreted proteins would be purified via Nickel affinity chromatography [9] followed by size exclusion chromatography. Note that His-tags were included within each of our constructs.

Figure 3. Protein purification protocol.

Section 4. Assay protocols

We proposed to conduct 3 assays for detecting and testing the enzymatic activities of PETase-MHETase and MHETase-PETase fusion proteins, and also PETase and MHETase on their own.

Figure 4. Assays for testing the enzymatic activity of PETase, MHETase, PETase-MHETase fusion and MHETase-PETase fusion proteins.
Assay 1: para-nitrophenol-acetate (pNPA) assay. Assay 2: Western Blot assay for the 4 constructs in which samples would be loaded on lane 3-6 while lane 1 and 2 would be loaded with blank control and ladder respectively. Assay 3. HPLC assay for enzymes identification test as well as quantitatively analysis of enzymatic activities.

Assay 1: The colorimetric para-nitrophenol-acetate (pNPA) assay: pNPA would be decomposed into p-nitrophenol (pNP) and acetate with the presence of active PETase [10]. The reaction product pNP is visually detectable with the sample becoming yellow. The enzymatic activity of PETase can be quantified by testing the absorbance of pNP at 405nm. According to the results from last year’s Exeter iGEM team, it was shown that the relationship between PETase enzymatic activity and pNPA substrate concentration is linear with positive gradients. 10mM was the maximum substrate concentration Exeter used to test to understand the extent of the PETase enzymatic activity, however, we proposed to elevate the level of substrate concentration to make it beyond 10mM. This would finally give us a flat line as the result of saturation. Knowing the saturation point could help us identify what would be the restricting factor and most appropriate substrate concentration for maximizing bio-electricity production because TPA produced from the enzymatic reaction would be further utilized by P. putida to convert into lactate. Since 3 of the 4 constructs contain the element of PETase, pNPA assay would be done on constructs 1-3.

Assay 2: For the Western Blot assay [12], we have used Bioinformatics Protein Molecular Weight by uploading our amino acid sequence onto it to predict the molecular weights of the fusion proteins which should be approximately 215kDa. The molecular weights of PETase and MHETase were identified in literature as 30kDa and 65kDa, respectively. Lane 1 would be an empty vector plasmid as the negative control.

Assay 3: The HPLC assay [13] would be used for testing the enzymatic activities of all 4 constructs. We would use PET film and MHET as the substrates to test the enzymatic activities of PETase and MHETase respectively. The major product of PETase activity should be MHET with a small proportion of the by-product, BHET while the major product of MHETase activity should be TPA with a relatively small proportion of ethylene glycol (EG).

For testing PETase in constructs 1, 2 and 3, after PET film is supplied to the samples, we would detect and quantify the concentration of MHET and the by-product, BHET by comparing our results with MHET and BHET HPLC standard curves and integrating the area under the peaks. The results from testing the fusion proteins could also inform us of whether the enzymatic activity of PETase decreases when compared with the result of testing single PETase.

For testing MHETase in constructs 1, 2 and 4, MHET would be supplied to the samples and the concentration of TPA would be integrated. Likewise, the result of construct 4 (single MHETase) can be compared with the results from constructs 1 and 2 to see whether MHETase enzymatic activity decreases in the fusions.

However, there would be a possibility that when testing the activity of PETase in the fusion proteins, (i.e. constructs 1 and 2) we might not be able to detect MHET if it would be converted to TPA extremely quickly. In this case, we would apply a continuous HPLC monitoring or taking samples at much shorter intervals. All results would be compared with the MHET, BHET and TPA standard curves to confirm the formation of the metabolites.

Originally, we planned to send Exeter our constructs synthesized by Genscript; however, our order was delayed and still being processed unfortunately. As a result, we do not have the wet lab results this year.

Objective 2: Expressing TPA transporter and degradation pathway in P. putida

The products generated by the fusion protein are Ethylene glycol (EG) and Terephthalic acid (TPA). Our second objective was to engineer P. putida to metabolise TPA fully and channel the carbon towards production of lactate given its metabolic versatility [14].

From literature research, we found that P. putida cannot degrade TPA; however, P. putida makes use of the Beta-ketoadipate pathway [15] to degrade the aromatic compound Protocatechuate (PCA). Comamonas Testosteroni was shown to convert TPA to PCA via the intermediate 1,2-dihydroxy-3,5-cyclohexadiene-1,4-dicarboxylate (DCD) [16]. There is also literature showing successful expression of these enzymes in P. putida using a broad host range plasmid [17] [18]. Therefore, our goal was to express them in P. putida to convert TPA to PCA. This required expression of both the TPA transporter [18] (Fig. 1) and the catabolic operon [16] (Fig. 2) from C. testosteroni in P. putida.

Figure 1.Components of the TPA transporter.

The TPA transporter (Fig. 1) consists of the tpiA major subunit and the tpiB minor subunit, both of which are transmembrane proteins. TphC is the receptor that senses the TPA present in the environment; however, this is encoded within the catabolic operon.

Figure 2. TPA catabolic operon.

The catabolic operon consists of the TphC receptor, a two subunit TPA dioxygenase (TPADO) encoded in genes TphA2 and TphA3, a reductase component of TPADO encoded in TphA1 and the DCD dehydrogenase encoded in TphB. The oxygenase/reductase components of TPADO are involved in converting TPA to DCD, while DCD dehydrogenase yields PCA. PCA can then be naturally degraded by P. putida using the Beta-ketoadipate pathway. Genes tphA2, tphA3 and tphB are under the the same RBS because the genes have small regions of overlap whereby the beginning of the next gene starts before the previous one has reached the stop codon. This gene structure allows the TPADO to fold into a large multisubunit complex.

We decided to synthesise the TPA transporter and the catabolic operon as contigs. These 2 parts were assembled into a broad host range plasmid (pBBR1k) that has been previously tested in P. putida [17] using Gibson Assembly [19]. For this, specific primers with overhangs had to be designed in order to ensure efficient overlapping of the single stranded regions, thereby directing the assembly. The final assembly had the following characteristics:

Primers had 40 to 60% GC content with the 3´ ending with G or C to promote binding
Primers Length were between 20 and 30 bp
Primers with melting temperatures (Tm) ranging from 50°C to 65°C and with 5°C Tm difference
Assembly was codon optimized for P. putida
Both the TPA transporter subunits and the TPA catabolic operon were synthesised from GenScript

Lab experiments will be performed by next year´s UCL iGEM team, but we have provided the following potential experimental protocols on P. putida that perhaps can act as the reference for the next year team.

Section 1. Cloning and assembly of TPA transporter and catabolic operon

We have transferred the sequence of the genes of TPA transporter and catabolic operon from literature onto Benchling before conducting codon optimization for the genes in order to comply with RFC10. On Benchling, we assembled TPA transporter, catabolic operon and linearized plasmid pBBR1K using Gibson Assembly.

Link to this construct design on Benchling: TPA transportor + Catabolic operon (Gibson assembly).

Figure 1. TPA transporter and catabolic operon cloning mechanism.
TPA transporter, catabolic operon and linearized plasmid pBBR1K would be assembled via Gibson Assembly.

Section 2. Transformation and P. putida cell growth protocols

We proposed to use P. putida KT2440 as expression host. During the initial process of cell culture, we planned to supply glucose to ensure we would have sufficient number of cells before switching to give TPA as the only food source.

Figure 2. Proposed transformation of plasmid pBBR1K assembled with TPA transporter and catabolic operon into P. putida KT2440.

Section 3. TPA transporter and catabolic operon testing and lactate production assay

After sufficient number of cells have been produced on glucose, we proposed to switch the growth medium to a defined medium containing TPA only. We would monitor the process of cell growth by taking samples periodically and measuring the samples’ optical density at 600nm. If biomass increases, we could confirm that TPA transporter is correctly expressed on the cell membranes, and subsequently, conduct HPLC to measure the level of lactate in the media so that we could confirm TPA catabolic operon works. A decrease in biomass may indicate that either TPA cannot be utilized by P. Putida or the level of cell mass growth on TPA is not equivalent to that on glucose, we would conduct HPLC to measure the level of TPA in the samples. If the concentration of TPA ([TPA]) decreases, we could confirm that TPA transporter works well whereas there might be problems with catabolic operon, resulting in the decrease in biomass. However, if [TPA] does not decease, TPA transporter could have potential problems, but whether the catabolic operon works remains unknown.

Figure 3. Assays for testing the functionality of TPA transporter and catabolic operon using spectroscopy and HPLC respectively.

Part 2: Coculture Design

Overall Aim

Integrating the work from the Modelling and Synthetic Biology groups, we concentrated on tackling the oxygen problem of coculturing E. coli, P. putida, and S. oneidensis.

We learned that within our coculture, S. oneidensis, should be cultured anaerobically since it is well known that O2 acts as an electron acceptor, thereby minimising the current output of any MDC [20]. Further research and communications with Dr. Christy Dykstra and Tyler Myers from University of California San Diego (UCSD) informed us that E. coli BL21 (DE3) and a mutant strain of P. putida could be cultured under anaerobic conditions.

Based on initial research, we proposed to coculture three species together in one anaerobic chamber as illustrated in the diagram below.

Figure 1. Overview of the mechanism of the co-culture.

E. coli feeds on PET by secreting the PETase-MHETase fusion to the extracellular environment generating Ethylene glycol (EG) and TPA. Ethylene glycol is used to feed E. coli, while TPA feeds P. putida in order to maximise secretion of Lactate. The lactate generated can then feed S. oneidensis, which grows as a biofilm on the anode surface and generates a current output.

We performed Flux Balance Analysis (FBA) [22] to investigate the impact of E. coli and P. putida ’s oxygen uptake level on their biomass growth and lactate production. We simulated two scenarios for the coculture: anaerobic culture and aerobic culture.

For the details of FBA, please visit our Model page.

Results of FBA simulations are as demonstrated in the figure below.

Figure 1. Comparison of the growth of the co-culture under aerobic and anaerobic conditions.
The upper panels represent our FBA simulation results: E. coli and P. putida grow better under aerobic conditions and thereby generate good amounts of lactate. The lower panels represent what we learned from literature research with regards to S. oneidensis: a higher current is generated under anaerobic conditions with electrons being transferred to the anode. However, under aerobic conditions, oxygen acts as the electron acceptor and a lower current output is generated.

Analyses:

Anaerobic culture: without oxygen, E. coli is not able to degrade PET and grow on EG; therefore, no TPA is produced to support P. putida growth. Nevertheless, since E. coli is a facultative anaerobe, if we supply glucose, E. coli can still achieve a lower level of biomass growth and lactate production through its native fermentation.
Aerobic: with enough oxygen supplied, E. coli and P. putida can grow successfully and produce TPA and lactate, respectively. In our simulation, P. putida can produce up to 8. 2131 mmol/gDCW/h, sufficient to support the biomass growth of S. oneidensis.

From FBA simulation results, we have concluded that aerobic condition is needed for the two plastic degrading species, E. coli and P. putida to achieve our target level of plastic degradation and produce lactate to support S. oneidensis growth.

To tackle the challenge of oxygen availability, we have altered our original design of coculturing all three species to coculture E. coli and P. putida in an aerobic chamber and S. oneidensis in an anaerobic chamber purged with nitrogen [24].

Part 3: Microbial Desalination Cell (MDC) Operation and Configuration Design

Overall Aim

With the combined efforts from modeling and human practices groups, we aimed to propose a MDC configuration that is suitable for our coculture design mentioned in section 2 and maximizes desalination efficiency.

From literature review, we learned about the simplest configuration of MDC, the traditional H-shaped design including an anode chamber, a saline water chamber, and a cathode chamber [25].

To accommodate to our initial coculture design of culturing E. coli, P. putida, and S. oneidensis all in one anode chamber, we proposed that the following H-shaped configuration and operation design should be able to provide us with some preliminary results [26].

Figure 1. Initial proposal of MDC operations and configuration design.

Due to the change in our coculture design following the FBA simulations, it was necessary that we changed our MDC configuration and operation designs accordingly. Through communications with Christy and Tyler from UCSD, we learned to adopt a 2-step process for plastic degradation considering that lactate produced by P. putida in the aerobic chamber is required to transfer to the anaerobic anode chamber to feed exoelectrogen S. oneidensis. Regarding configuration designs, we carried out further literature research and learned that Stacked Design and Concentric tubular Design have been proved to achieve the highest desalination efficiency, approximately 95% and above. Nevertheless, due to the challenges of constructing a Concentric tubular MDC, Christy suggested that we start our experimentations with simpler configurations at the early stages.

Incorporating advice from Tyler and Christy as well as our literature research, we have modified our MDC designs and proposed the following for operations and configurations [26]:

1. Operation design: Three-step process

Figure 1. Three-step process for MDC operations.

Step 1: E. coli and P. putida are cocultured aerobically to degrade PET and produce lactate.
Step 2: Centrifugation is performed for cell debris collection.
Step 3: Lactate is used for feeding S. oneidensis to generate bioelectricity separating sodium and chloride ions, thereby achieving desalination.

2. Configuration design: Three-stage development

Stage 1: H-shaped Design

Figure 2. H-shape design.

This figure presents the H-shaped configuration proposed by Tyler and Christy for our initial stage of MDC experiments as this configuration is the simplest to construct and operate.

Stage 2: Stacked Design

With the experience from stage 1, we will aim to improve desalination efficiency in stage 2; therefore, we are proposing a stacked design.

Figure 3. Stacked Design.
This illustration demonstrates our proposed stacked design in which S. oneidensis will grow in the anode chamber. This design includes multiple Ion Exchange Membranes (IEMs): Anion Exchange Membranes (AEMs) and Cation Exchange Membranes (CEMs). In this figure, anions are chloride ions while cations are sodium ions.

By adopting a series of alternating IEMs, we can increase desalination efficiency drastically as the number of ion pairs separated per electron transferred is increased. As illustrated in the figure above, if we were to install three pairs of anode and cathode exchange membranes in our configurations, every single electron transfer can separate three pairs of sodium and chloride ions [25].

Stage 3: Concentric tubular Design

Figure 4. Concentric tubular design.

To maximize desalination efficiency, we proposed this concentric tubular design where the distance between anode and cathode is minimized and the surface area for S. oneidensis biofilm is maximized. As the construction of this configuration is very challenging, we suggest to only perform experimentations when sufficient trouble-shooting experience has been acquired from Stage 1 and 2.

Conclusion

To sum up, we optimized our construct design, coculture design as well as MDC operation and configuration design. As a result of the construct design, we proposed to achieve bacterial plastic degradation by engineering a MHETase-PETase fusion protein in E. coli together with a TPA transporter-catabolic operon construct in P. putida. For the coculture design, we suggested to coculture E. coli and P. putida in a separate aerobe chamber while S. oneidensis to be cultured anaerobically in the anode chamber of MDC. Finally, to accommodate to the coculture strategy, we modified our MDC designs, currently comprised of three operation steps and three configuration development stages.

During the process of developing these three key aspects of our project design, we went through the design cycle iteratively and incorporated the various divisions of our project including Synthetic Biology, Modeling, Human Practices, and Collaborations. On this page, we have presented how these divisions were weaved into PETZAP, building up our solutions to tackle plastic pollution and freshwater scarcity. For the complete stories of these divisions, please visit Parts, Model, Human Practices, Collaborations, and Partnership.

Yoshida S, Hiraga K, Takehana T, Taniguchi I, Yamaji H, Maeda Y, et al. A bacterium that degrades and assimilates poly(ethylene terephthalate). Science (80- ). 2016 Mar 11;351(6278):1196–9.
Seo H, Kim S, Son HF, Sagong HY, Joo S, Kim KJ. Production of extracellular PETase from Ideonella sakaiensis using sec-dependent signal peptides in E. coli. Biochem Biophys Res Commun. 2019 Jan 1;508(1):250–5.
Sagong HY, Seo H, Kim T, Son HF, Joo S, Lee SH, et al. Decomposition of the PET Film by MHETase Using Exo-PETase Function. ACS Catal [Internet]. 2020 Apr 17 [cited 2020 Oct 26];10(8):4805–12. Available from: https://dx.doi.org/10.1021/acscatal.9b05604
Kelley LA, Mezulis S, Yates CM, Wass MN, Sternberg MJE. The Phyre2 web portal for protein modeling, prediction and analysis. Nat Protoc [Internet]. 2015 Jun 30 [cited 2020 Oct 26];10(6):845–58. Available from: http://predictioncenter.org/index.cgi
Knott BC, Erickson E, Allen MD, Gado JE, Graham R, Kearns FL, et al. Characterization and engineering of a two-enzyme system for plastics depolymerization. Proc Natl Acad Sci [Internet]. 2020 Sep 28 [cited 2020 Oct 26];117(41):202006753. Available from: www.pnas.org/cgi/doi/10.1073/pnas.2006753117
Beckwith J. The Sec-dependent pathway. Res Microbiol. 2013 Jul 1;164(6):497–504.
Almagro Armenteros JJ, Tsirigos KD, Sønderby CK, Petersen TN, Winther O, Brunak S, et al. SignalP 5.0 improves signal peptide predictions using deep neural networks. Nat Biotechnol [Internet]. 2019 Apr 1 [cited 2020 Oct 26];37(4):420–3. Available from: https://www.nature.com/articles/s41587-019-0036-z
Klein JS, Jiang S, Galimidi RP, Keeffe JR, Bjorkman PJ, Regan L. Design and characterization of structured protein linkers with differing flexibilities. In: Protein Engineering, Design and Selection [Internet]. Oxford University Press; 2014 [cited 2020 Oct 26]. p. 325–30. Available from: https://academic.oup.com/peds/article/27/10/325/1455209
Block H, Maertens B, Spriestersbach A, Brinker N, Kubicek J, Fabis R, et al. Chapter 27 Immobilized-Metal Affinity Chromatography (IMAC). A Review. In: Methods in Enzymology. Academic Press Inc.; 2009. p. 439–73.
Ma Y, Yao M, Li B, Ding M, He B, Chen S, et al. Enhanced Poly(ethylene terephthalate) Hydrolase Activity by Protein Engineering. Engineering. 2018 Dec 1;4(6):888–93.
Shetty RP, Endy D, Knight TF. Engineering BioBrick vectors from BioBrick parts. J Biol Eng [Internet]. 2008 [cited 2019 Nov 2];2(1):5. Available from: http://jbioleng.biomedcentral.com/articles/10.1186/1754-1611-2-5
Size Exclusion Chromatography of Protein Biopharmaceuticals: Past, Present and Future | American Pharmaceutical Review - The Review of American Pharmaceutical Business & Technology [Internet]. [cited 2020 May 10]. Available from: https://www.americanpharmaceuticalreview.com/Featured-Articles/348912-Size-Exclusion-Chromatography-of-Protein-Biopharmaceuticals-Past-Present-and-Future/
Kaufmann SH, Ewing CM, Shaper JH. The erasable Western blot. Anal Biochem. 1987 Feb 15;161(1):89–95.
Ismail BP. Basic Principles of Chromatography. In Springer, Cham; 2017. p. 185–211.
Johnson CW, Beckham GT. Aromatic catabolic pathway selection for optimal production of pyruvate and lactate from lignin. Metab Eng. 2015 Feb 1;28:240–7.
Stainer RY, Ornston LN. The β-Ketoadipate Pathway. Adv Microb Physiol. 1973 Jan 1;9(C):89–151.
Sasoh M, Masai E, Ishibashi S, Hara H, Kamimura N, Miyauchi K, et al. Characterization of the Terephthalate Degradation Genes of Comamonas sp. Strain E6. Appl Environ Microbiol [Internet]. 2006 [cited 2020 Apr 28];72(3):1825–32. Available from: http://aem.asm.org/
Cook TB, Rand JM, Nurani W, Courtney DK, Liu SA, Pfleger BF. Genetic tools for reliable gene expression and recombineering in Pseudomonas putida. J Ind Microbiol Biotechnol [Internet]. 2018 Jul 1 [cited 2020 Oct 26];45(7):517–27. Available from: /pmc/articles/PMC6161825/?report=abstract
Hosaka M, Kamimura N, Toribami S, Mori K, Kasai D, Fukuda M, et al. Novel tripartite aromatic acid transporter essential for terephthalate uptake in Comamonas sp. strain e6. Appl Environ Microbiol. 2013;79(19):6148–55.
Gibson DG, Young L, Chuang RY, Venter JC, Hutchison CA, Smith HO. Enzymatic assembly of DNA molecules up to several hundred kilobases. Nat Methods [Internet]. 2009 Apr 12 [cited 2020 Oct 19];6(5):343–5. Available from: http://npg.nature.com/reprintsandpermissions/
Gibson DG, Young L, Chuang RY, Venter JC, Hutchison CA, Smith HO. Enzymatic assembly of DNA molecules up to several hundred kilobases. Nat Methods [Internet]. 2009 Apr 12 [cited 2020 Oct 19];6(5):343–5. Available from: http://npg.nature.com/reprintsandpermissions/
Logan BE. Exoelectrogenic bacteria that power microbial fuel cells. Nat Rev Microbiol [Internet]. 2009 Mar 30 [cited 2020 Oct 26];7(5):375–81. Available from: https://www.nature.com/articles/nrmicro2113
Orth JD, Thiele I, Palsson BO. What is flux balance analysis? [Internet]. Vol. 28, Nature Biotechnology. Nature Publishing Group; 2010 [cited 2020 Oct 26]. p. 245–8. Available from: https://www.nature.com/articles/nbt.1614
Chen S, Patil SA, Brown RK, Schröder U. Strategies for optimizing the power output of microbial fuel cells: Transitioning from fundamental studies to practical implementation. Vols. 233–234, Applied Energy. Elsevier Ltd; 2019. p. 15–28.
Kim Y, Logan BE. Microbial desalination cells for energy production and desalination. Desalination. 2013 Jan 2;308:122–30.
Al-Mamun A, Ahmad W, Baawain MS, Khadem M, Dhar BR. A review of microbial desalination cell technology: Configurations, optimization and applications. Vol. 183, Journal of Cleaner Production. Elsevier Ltd; 2018. p. 458–80.