Introduction

We have achieved engineering success by repeating the DBTL (Design-Build-Test-Learn) cycle several times. DBTL is an essential engineering principle in achieving a zero-one (create from nothing) or improving on an existing one. It is the fundamental engineering principle that drives our project. Furthermore, Waseda-Tokyo's DBTL cycle was a successful DBTL cycle, which was reciprocated not only by the Wet Lab but also by the dry Lab. The trajectory of the project is documented below.

BIP Module

We selected BIND-bearPETase by creating a BIND-PETase using existing PETases and any PETase variants we created shown in Table.4.1.1. Specifically, we followed the cycle shown in Fig. 4.1.1 to verify and compare the enzyme activity of the PETases. This process was divided into three major phases: in Phase 1, BIND-PETase (WT, E. coli optimized, 5 mut.) was designed and its basic function was verified; in Phase 2, duraPETase ^[1], which was also created based on evolutionary engineering, was used to create BIND-duraPETase. In addition, machine learning and rationalization were used to develop the duraPETase. In Phase 3, we investigated the BIND-PETase by detailed functional analysis. Multiple DBTL cycles were being rotated in each phase.

Fig. 4.1.1. DBTL cycle flow in BIP Module

Table. 4.1.1. List of BIND-PETase variants

Taxonomy	BIND-PETase variant	Description
Existing variants	WT	Wild type sequence from I. Sakaiensis
	E. coli optimized	WT codon optimized for E. coli
	5 mut.	5 amino acid substitution sequence of E. coli optimized
	dura	10 variants of WT, with superior activity and heat tolerance
iePETase	ID1	ESM_PETase
	ID2	ESM_PETase_affinity_screenig
	ID3	ESM_PETaseaffinity_REU_screening
	ID4	DuraPETase_EvoProGrad_EMS2_evocouplings_affinity_screening
	ID5	DepoPETase_EvoProGrad_EMS2_evocouplings_affinity_screening
rdPETase	ID6	LCC_mut_IsPETase_M1M3
	ID7	LCC_mut_IsPETase_M3
	ID8	LCC_mut_IsPETase_M1M2M3
	ID22	DepoPETase-M4_DuraPETase-M10
	ID23	DepoPETase-M2-3_DuraPETase-M10
	ID24	DepoPETase-M2_DuraPETase-M10
	ID25	DepoPETase-M3_DuraPETase-M10
	ID26	DepoPETase-M2-2_DuraPETase-M10
	ID27	DepoPETase-M2-1_DuraPETase-M10

Phase 1. Construction of BIND-PETase (WT, E. coli optimized, 5 mut.) and Characterization of Functionality

We first created BIND-PETase(WT) by fusing an existing PETase(WT) (BBa_ J428065) from I. sakaiensis with CsgA, a factor in the formation of E. coli curli fibers. The detailed name of the plasmid is pMAL-c4X-RBS-csgA-BamHI_linker-PETase(WT) and its abbreviation is BIND-PETase(WT) in the following.

In addition to the PETase(WT), we also created BIND-PETase with several variants created in the past. Among the various mutant PETases, we used two variants that were registered in iGEM parts and were easily accessible to create further BIND-PETases. One is a PETase (E. coli optimized) (BBa_K2910000), in which the codon of PETase (WT) is optimized for E. coli, and the other is a PETase (5 mut.) (BBa_K2910003), in which the codon is optimized and has 5 mutations.
　Through these following DBTL cycles, we have succeeded in construction of each plasmid which encodes BIND-PETase(WT, E. coli optimized, 5 mut.). Furthermore, we were able to confirm the activity of these BIND-PETases and the curli fiber formation ability, which will be explained later.

1.1 Construction and expression characterization of BIND-PETase(WT)

We constructed the plasmid encoding BIND-PETase (WT) using the GenSynthesis service provided by GenScript, based on the previous research by Zhu B. et al. ^[2]. Although they constructed the plasmid encoding BIND-PETase (WT), the details were not fully described. While information regarding CsgA and PETase (WT) was provided, the rest was not specified, requiring us to construct it ourselves. The following outlines the process according to the DBTL cycle.

≪cycle 1-1-1: optimal RBS design for BINDPETase (WT) expression≫

Design

To efficiently drive the BIND-PETase module, an optimized RBS is essential. Without the appropriate RBS strength, overexpression could overload the E. coli BL21(DE3)
, while insufficient expression might result in no activity. We needed to optimize the RBS to regulate the amount of CsgA displayed on the surface of E. coli BL21(DE3) ensuring it remains at an appropriate level.

Build

Therefore, we referred to previous research and calculated the RBS efficiency required for appropriate CsgA expression using Salis Lab's RBS Calculator. We examined the RBS strength of the part “pRha + CsgA” (BBa_K1583100), which allows for Rhamnose promoter-dependent overexpression of CsgA, to evaluate whether it could be incorporated into our plasmid.

Test

The transcriptional rate of the RBS attached to pRha + CsgA was calculated by RBS caluculator and found to be 40.80. First, the Promoter Calculator from Salis Lab was used to identify the transcription start point (+1), and then the RBS Calculator was utilized to determine the intensity downstream of that point. The results are shown in Fig. 4.1.2.

Fig. 4.1.2. Transcriptional Rate of RBS attached to pRha+CsgA(BBa_K1583100)

Learn

The intensity of 40.80 was considered a very low transcriptional rate based on our experience. Consequently, we concluded that the RBS attached to pRha + CsgA could not achieve our desired expression level of CsgA. Therefore, in the next Cycle 1-1-2, we decided to conduct a separate investigation into the RBS strength of CsgA used in biotypic catalysts, such as BIND-PETase, from previous studies.

≪cycle 1-1-2: optimal RBS design for BIND-PETase (WT) expression≫

Design

Based on the reflections from Cycle 1-1-2, we reviewed previous research focused on the overexpression of CsgA for the construction of a whole-cell biocatalyst. As a result, we decided to refer to the plasmid Rec-PhoA/CsgA (Addgene #170787) ^[3].

Build

We obtained the sequence for Rec-PhoA/CsgA.

Test

The transcriptional rate of its RBS was calculated to be 699.39. As in the Test of Cycle 1-1-2, the Promoter Calculator was used to identify the transcription start point (+1) for Rec-PhoA/CsgA, followed by the RBS Calculator to determine the transcriptional rate. The output from the RBS calculator is shown in Fig. 4.1.3.

Fig. 4.1.3. Transcriptional rate of RBS for CsgA expression included in pRha+CsgA

Learn

Based on the Test, we were able to determine the target value for the transcriptional rate for BIND-PETase. This value of 699.39 was greater than that of the RBS attached to pRha + CsgA, which ranged from 500 to 1000. Therefore, we considered designing an RBS with a similar transcriptional rate for the BIND-PETase sequence in Cycle 1-1-3.

≪cycle 1-1-3: optimal RBS design for BIND-PETase (WT) expression≫

Design

From the Cycle1-1-2 efforts, the transcriptional rate is considered to be between 500 and 1000 for driving the BIND-PETase Module efficiently.

Build

Therefore, RBS sequences were generated by setting the transcriptional rates to 600, 700, 800, and 900 using the RBS Calculator for the BIND-PETase (WT) sequence, which was designed based on previous studies.

Test

After examining the generated RBSs, we found that the optimal RBS sequence was CGGTCGGTTCCGGGGGCCCTATTCAT, which has a transcriptional rate of 800 and exhibits no significant bias toward any particular base. The RBS sequence and simulation results are shown in Fig. 4.1.4. We adopted this RBS sequence and synthesized a vector for BIND-PETase (WT) expression using GenScript's GeneSynthesis service.

Fig. 4.1.4. Optimized RBS for BIND-System generated with Transcriptional Rate set to 800

Learn

We found that the transcriptional rate of the RBS we designed was appropriate. As shown in DBTL Cycles below, western blotting and functional analysis assays of various BIND-PETases show that this RBS works as we expected and induces expression of BIND-PETase in appropriate amounts.

Our New Composite Parts, RBS+BIND-bearPETase (BBa_K5436124), is registered including this RBS. We also registered the RBS by itself as Basic Parts (BBa_5436005) and contributed to the iGEM Parts Library.

≪cycle 1-1-4: characterization of BIND-PETase (WT) expression≫

Design

In order to confirm whether BIND-PETase (WT) designed in cycle1-1-3 is expressed in E. coli BL21(DE3), SDS-PAGE was planned.

Build

BIND-PETase was transformed into E. coli BL21(DE3) and IPTG induction was performed.

Test

When E. coli BL21(DE3)was lysed and subjected to total protein SDS-PAGE, no differences were observed between the IPTG+ and IPTG- conditions. Specifically, there was no variation in the band corresponding to approximately 46.6 kDa ^[2], which is the molecular weight of BIND-PETase (WT) (refer to Results).

Fig. 4.1.5. The result of Total Protein SDS-PAGE

Samples derived from two colonies, each with IPTG+/-

Learn

The sample contains a wide variety of proteins, which may mask the presence of BIND-PETase. Therefore, we considered it necessary to specifically detect BIND-PETase (WT).

≪cycle 1-1-5: characterization of expression of BIND-PETase (WT)≫

Design

Western Blotting was planned to specifically detect BIND-PETase (WT).

Build

Since the C-terminus of BIND-PETase (WT) contains a His-tag, we decided to use a His-tag antibody as the primary antibody. Induction was performed on transformed E. coli BL21(DE3), as described in Cycle 1-1-3.

Test

Western blotting with the His-tag antibody clearly revealed a band around 45 kDa in the IPTG-induced sample (refer to Results).

Fig. 4.1.6. The result of Western Blotting

In the IPTG(+) sample, the His-tag antibody showed a prominent band around 45 kDa, which corresponds to the molecular weight of BIND-PETase. In contrast, no band was detected in the IPTG(-) sample.

Learn

The expression of BIND-PETase (WT) was confirmed by IPTG induction in E. coli BL21(DE3) transfected with BIND-PETase (WT), suggesting that the designed RBS was suitable for its expression. The reason BIND-PETase was not clearly detected by SDS-PAGE is likely due to its low expression level, which may have been masked by endogenous proteins. The protein present on the membrane has a low expression level^[4], which could be one reason why BIND-PETase (WT) could not be detected by SDS-PAGE.

1.2 Basic functional verification of BIND-PETase (WT, E. coli optimized, 5 mut.)

The results from the previous section confirmed the successful construction of the plasmid for expressing the fusion protein BIND-PETase (WT). We then decided to create two types of BIND-PETase: BIND-PETase (E. coli optimized)(BBa_K5436131) and BIND-PETase (5 mut. )(BBa_K5436132), and validate their ability for Curli fiber formation, as well as their enzymatic activity, including that of BIND-PETase (WT)(BBa_K5436130).

≪cycle 1-2-1: construction and functional validation of BIND-PETase (WT, E. coli optimized, 5 mut.) ≫

Design

To obtain the plasmids encoding BIND-PETase(E.coli optimized) and BIND-PETase(5 mut.), we considered replacing only the PETase portion of the BIND-PETase (WT) constructed in the previous section. Additionally, the display of BIND-PETase on the E. coli BL21(DE3) membrane surface was confirmed by the Congo Red Assay, which quantifies curli fiber formation. The enzyme activity is planned to be assessed through the hydrolysis of a molecule called para-Nitrophenyl Butyrate(pNPB).

Build

BIND-PETase (E. coli optimized) and BIND-PETase (5 mut.) were constructed using NEBuilder Assembly and introduced into E. coli BL21(DE3). For more detail, refer to the Experiments.

Test

The formation of curli fibers on the membrane surface was confirmed by the Congo Red Assay, whereas the activity of PETase was measured using the pNPB Assay.

Both the curli formation and hydrolysis activity were confirmed in BIND-PETase (WT) and BIND-PETase (E. coli optimized), but not in BIND-PETase (5 mut.). Please refer to the Experiments section for the assay protocol and the Results section for details.

Fig. 4.1.7. Congo Red assay results for BIND-PETase(WT, E. coli optimized, 5 mut.)

Fig. 4.1.8. pNPB assay results for BIND-PETase(WT, E. coli optimized, 5 mut.)

Learn

BIND-PETase (WT) and BIND-PETase (E. coli optimized) were found to exhibit curli fiber-forming ability and hydrolytic activity. In contrast, the BIND-PETase(5 mut.) did not function as intended. A possible reason for this failure could be that the CsgA and PETase (5 mut. ) domains were structurally incompatible, which may have hindered their polymerization and activity with each other.

Phase 2. Construction and functional characterization of BIND-PETase based on evolutionary engineering.

We aimed to enhance BIND-PETase through evolutionary engineering, building on the successful creation of BIND-PETase in Phase 1.

2.1 Characterization of BIND-duraPETase

DuraPETase, which is known for its high heat tolerance and activity, was created through directed evolution^[1] and possesses 10 mutations compared to PETase(WT). We constructed a plasmid encoding BIND-duraPETase and conducted the following functional analysis.

≪cycle 2-1-1: Construction and functional verification of BIND-duraPETase≫

Design

To obtain the plasmids encoding BIND-duraPETase, we planned to replace the PETase portion of BIND-PETase (WT) constructed in the previous section with duraPETase.

Build

Using NEBuilder Assembly, we constructed the plasmid encoding BIND-duraPETase. In this process, the PETase portion of BIND-PETase (WT), constructed in a previous cycle, was replaced with a duraPETase fragment synthesized through Twist Bioscience's Gene Fragments service. The sequence was confirmed to be correct and subsequently introduced into E. coli BL21(DE3).

For more detail, refer to Experiments.

Test

The Congo Red Assay and pNPB Assay confirmed that BIND-duraPETase exhibits curli fiber-forming ability and hydrolytic activity against pNPB (refer to Results).

Fig. 4.1.9. Congo Red assay results for BIND-PETase(WT, E. coli optimized, 5 mut., dura)

Fig. 4.1.10. pNPB assay results for BIND-PETase(WT, E. coli optimized, 5 mut., dura)

Learn

BIND-duraPETase was found to be more active than BIND-PETase(WT). This increased activity is attributed to the fact that duraPETase was designed through evolutionary engineering. This suggests that our BIND-PETase generation method can be applied to PETases developed through evolutionary engineering approaches.

2.2 In Silico Evolution Pipeline / Integrate Rational Design Pipeline for the Design of iePETase / rdPETase

We designed PETase variants using the In Silico Evolution Pipeline and Integrate Rational Design Pipeline, proposing 14 candidate variants for bearPETase. Given the vast search space of 20^300 possible combinations, machine learning and rational design were used to efficiently select 5 iePETase and 9 rdPETase candidates.

Fig. 4.1.11. The overview of In Silico Evolution Pipeline

14 new PETase variants, assigned IDs from ID1 to ID8 and ID22 to ID27. Of these, IDs 1 to 5 represent iePETase, while IDs 6 to 8 and 22 to 27 represent rdPETase. Our goal was to construct mutated BIND-PETase (BIND-iePETase, BIND-rdPETase) using these variants and select the optimal one. For the detailed sequence information, refer to Parts.

Among them, we identified the most active BIND-PETase variants. It is called BIND-bearPETase (BBa_K5436124) and registered as a New Composite Part. For the detailed characterization of this part, refer to BBa_K5436124.

The following section documents the process, following the DBTL cycle, through which BIND-PETase(ID24) was selected as BIND-bearPETase.

<<cycle 2-2-1: ESM3 (iePETase: ID1-3)>>

Design

In this cycle, we used the protein language model ESM3 ^[5] to design PETase variants. ESM3 can generate masked parts of amino acid sequences by inputting prompts related to amino acid sequences, coordinates, and functions. We used ESM3 to input information from a hydrolytic enzyme protein family, including binding and active site coordinates and sequences, generating 100 candidate PETase variants.

Build

Then, the 3D structures of proteins were generated using ESM3. Among the 100 candidates generated using ESM3, screening was conducted to select five candidates for ID1, five for ID2, and ten for ID3.

We selected the highest-scoring sequences for three parameters. ID1 referred to the reliability indicator of ESM3 (pTM), ID2 referred to the docking score from DockString ^[6], and ID3 referred to the Rosetta score ^[7] (REU).

Test

We performed docking simulation between PET and PETase using the molecular docking tool AutoDock Vina ^[8] and selected candidates for Wet experiments.

For ID1, after docking with PET monomers, we docked the best-bound PETase with PET trimmers. Among the results, the one with the lowest binding energy was selected as a candidate for Wet experiments. This decision was based on prior research.

As for ID2 and later IDs, criteria were established through discussions with experts, and the docking was conducted accordingly. For details on the method, refer to Model.

The results of molecular docking are as follows. In the figures, the binding sites are indicated by pink dots.

The binding energy of ID1 was -5.8 kcal/mol. The docking result is shown in Fig. 4.1.12.

Fig. 4.1.12. The result of molecular docking of ID1

The binding energy of ID2 was -6.8 kcal/mol. The docking result is shown in Fig. 4.1.13.

Fig. 4.1.13. The result of molecular docking of ID2

The binding energy of ID3 was -7.1 kcal/mol. The docking result is shown in Fig. 4.1.14.

Fig. 4.1.14. The result of molecular docking of ID3

Learn

Using the method described above, we proposed three PETase variants for IDs 1-3 for candidates. For ID1, two candidates with bound PET monomers were selected from five initial candidates. Then, we performed docking with PET trimers and proposed the one with the lowest energy as ID1. For ID2, three candidates with bound PET dimers were selected from five initial candidates. Docking was performed with PET tetramers, and the one where PET was bound in the “groove” near the binding site, as shown in the figure, was proposed as ID2. For ID3, four candidates with bound PET dimers were selected from ten initial candidates. After docking with PET tetramers, the one where PET was bound in the “groove” near the binding site, as shown in the figure, was proposed as ID3.

<<cycle 2-2-2: EvoPlotGrad (iePETase: ID4-5)>>

Design

In this cycle, we used a tool called EvoPlotGrad ^[9] to design PETase variants. EvoPlotGrad combines the output of various tools and sets the range considered to have good results, searching for candidates within the common region of those outputs. Using the output of ESM2 ^[10] (protein feature data) and EvCouplings (co-evolutionary feature data), EvoPlotGrad generated amino acid sequences for 30 PETase variants. ESM2 ensures the naturalness of the protein, and EvCouplings ^[11] ensures evolutionary naturalness. Furthermore, based on the docking scores from DockString, we selected the top ten candidates. ID4 used duraPETase as the starting sequence, and ID5 used depoPETase ^[12] as the starting sequence. This is how we obtained the amino acid sequences of the PETase variants.

Build

We generated the 3D structures of proteins using ESM3. Although the accuracy is lower compared to AlphaFold 2 ^[13], the processing speed was prioritized.

Test

As in cycle 2-2-1, we performed docking simulation between PET and PETase using AutoDock Vina and selected candidates for Wet experiments. We proposed one candidate each for ID4 and ID5.

The results of molecular docking are as follows. In the figures, the binding sites are indicated by pink dots.

The binding energy of ID4 was -4.7 kcal/mol. The docking result is shown in Fig. 4.1.15.

Fig. 4.1.15. The result of molecular docking of ID4

The binding energy of ID5 was -5.6 kcal/mol. The docking result is shown in Fig. 4.1.16

Fig. 4.1.16. The result of molecular docking of ID5

Learn

L: Using the method described above, we proposed two PETase variants for IDs 4-5. For ID4, three candidates with bound PET dimers were selected from ten initial candidates. Docking was performed with PET tetramers, but only one candidate was bound, which was proposed as ID4. As shown in the figure, PET was not bound well, so its activity in Wet experiments is not expected to be high. For ID5, five candidates with bound PET dimers were selected from ten initial candidates. Docking was performed with PET tetramers, and two candidates showed binding. Since both had the same binding energy in the tetramer docking, the one with lower energy in dimer docking was proposed as ID5.

<<cycle 2-2-3: LCC mut PETase (rdPETase: ID6-8)>>

Design

In this cycle, PETase variants were designed using the existing PET-degrading enzyme LCC ^[14]. LCC has higher activity toward PET than PETase, and its structure and sequence are very similar. We aligned the sequences and structures of LCC and PETase and identified three mutations in LCC that are expected to improve enzyme activity. These mutations were labeled M1, M2, and M3, and we designed seven PETase variants by combining these three mutations.

Build

We generated the 3D structures of proteins using AlphaFold 2.

Test

As in cycle 2-2-1, we performed docking simulation between PET and PETase using AutoDock Vina and selected candidates for Wet experiments. Based on the binding energy and docking pose, ID6 was M1M3, ID7 was M1, and ID8 was M1M2M3.

The results of molecular docking are as follows. In the figures, the binding sites are indicated by pink dots.

The binding energy of ID6 was -3.8 kcal/mol. The docking result is shown in Fig. 4.1.17.

Fig. 4.1.17. The result of molecular docking of ID6

The binding energy of ID7 was -4.8 kcal/mol. The docking result is shown in Fig. 4.1.18.

Fig. 4.1.18. The result of molecular docking of ID7

ID8 did not succeed in molecular docking. The docking result is shown in Fig. 4.1.19.

Fig. 4.1.19. The result of molecular docking of ID8

Learn

Using the method described above, we proposed three PETase variants for IDs 6-8. Out of the four candidates where PET bound in dimer docking, two candidates showed binding in tetramer docking. ID6 was proposed as M1M3 based on the low energy in dimer docking and the binding of PET near the "groove" in tetramer docking. ID7 was proposed as M1, and ID8, combining M1, M2, and M3, was proposed.

<<cycle 2-2-4: duraPETase × depoPETase(rdPETase : ID22-27)>>

Design

In this cycle, PETase variants were designed using depoPETase and duraPETase, which are existing PETase variants and known to have high enzymatic activity. These variants have enhanced enzyme activity due to different factors, and combining them is expected to further enhance enzyme activity.

Therefore, based on previous studies on depoPETase and duraPETase, we combined intermediate variants whose enzyme activity was evaluated by Wet experiments. Sharing mutation sites, Q119Y and Q119R, were removed to create the final 81 combinations.

Build

Next, AlphaFold 2 was used to generate the 3D structure of the protein due to its high accuracy, as it is crucial to examine hydrogen bonds within the 3D structure.

Test

We then visually inspected the order of mutations in PyMOL^[15] and used the presence or absence of unnatural hydrogen bond formation, as well as the position of hydrogen bonds in the active site, as key indicators. The six variants that performed well based on these criteria were designated as ID22-27.

Learn

L: Since the combination including M5 and after of duraPETase had an unnatural structure, M10, which has the highest enzymatic activity of duraPETase, and M4 and before of depoPETase were further analyzed. The results showed that

PETase(ID22): depoPETase-M4_duraPETase-M10
PETase(ID23): depoPETase-M2-3_duraPETase-M10
PETase(ID24): depoPETase-M2_duraPETase-M10
PETase(ID25): depoPETase-M3_duraPETase-M10
PETase(ID26): depoPETase-M2-2_duraPETase-M10
PETase(ID27): depoPETase-M2-1_duraPETase-M10

were selected.

2.3 Characterization of BIND-iePETase using designed iePETase / rdPETase

The next important question was whether the designed iePETase/rdPEtase actually constitutes and functions as a BIND-iePETase/rdPETase. There were many unknowns about the BIND-iePETase/rdPETase using our iePETase/rdPETase design. Therefore, it is important to verify the presence or absence of their activity in comparison with the BIND-PETase (WT, E. coli optimized) and BIND-duraPETase that have already been validated in this phase.

≪cycle 2-3-1: Construction and functional verification of BIND-iePETase/rdPETase ≫

Design

The backbone sequence including CsgA was adopted from BIND-PETase (WT), and only the PETase part was replaced with iePETase/rdPETase.

Build

Various iePETase/rdPETase sequences synthesized by Genscript, Twist bioscience and IDT were assembled with the backbone of BIND-PETase (WT) containing CsgA to obtain plasmids expressing BIND-iePETase/rdPETase. BIND-iePETase/rdPETase expression plasmid was obtained. For the detailed procedure, please refer to Experiments. These plasmids were transformed into E. coli BL21(DE3).

Test

The BIND-iePETase/rdPETase was tested for its ability to form Curli Fiber and hydrolyze ester bonds by the Congo Red Assay and pNPB Assay. BIND-PETase (ID3, 22~27) showed both curli fiber formation ability and hydrolytic activity. In particular, BIND-PETase (ID23~27) showed particularly high hydrolytic activity. On the other hand, BIND-PETase (ID1,2, 4~6) showed no or low curli fiber formation ability and hydrolytic activity.

Fig. 4.1.20. The result of Congo Red Assay

Fig. 4.1.21. The result of pNPB Assay

Learn

Among all the validated BIND-PETases, BIND-PETase (WT, E. coli optimized, dura, ID3, 22–27) demonstrated confirmed CsgA polymerization and PETase activity through the Congo Red Assay and pNPB Assay, which are essential functions of BIND-PETase. As a result, it progressed to Phase 3 for further functional validation. However, BIND-PETase (ID25-27), whose basic functions were confirmed, could not be validated further in Phase 3 due to time constraints. Additionally, BIND-PETases (5 mut., ID1, 2, 4–6), which did not demonstrate functionality as BIND-PETases, were not pursued for further validation.

Phase 3. Selection of superior BIND-PETase by further functional analysis

In this phase, we compared BIND-PETase, whose basic function was confirmed in phase 2, with BIND-PETase's practical advantages such as storage activity, reusability, and PET degradation activity.

≪cycle 3-1-1: Storage Activity Assay≫.

Design

To verify the societal applicability of BIND-PETases, we planned to measure their activity after storing them for a few weeks.

Build

BIND-PETases were stored in PBS suspension for 11 days at 4°C and room temperature (RT), shielded from light.

Test

The pNPB Assay was repeated on days 5 and 11, and the results were compared to those from the pNPB Assay conducted on day 0 of expression to examine their activity (refer to Results).

Fig.4.1.22. Storage Activity in the condition of room temperature.

Fig.4.1.23. Storage Activity in the condition of 4°C.

Learn

Storage Activity varied with mutant and storage temperature. Those with decreased activity due to storage may be due to degradation of BIND-PETase or decreased structural stability. Storage activity is important for the implementation of BIND-PETase and is an essential factor in the selection of BIND-bearPETase.

Furthermore, the increase in activity for ID23 and ID24 can be considered an advantage for the societal implementation of BIND-PETase. Since BIND-PETase functions as a whole-cell biocatalyst that is used in a living state, it is possible that cell proliferation and protein expression occurred during the storage period. This was an unexpected result for us, as we anticipated that activity would decrease over time, and we believe it could provide an advantage in societal implementation.

≪cycle 3-2-1: Reusability Assay≫

Design

D: We aim to measure whether BIND-PETase can be collected and reused after being utilized in a reaction, and to investigate how much its reusability varies among different mutants.

Build

The pNPB Assay was conducted using BIND-PETase, which was collected by centrifugation after the reaction. In this assay, the absorbance of pNP, produced by the hydrolysis of pNPB by PETase, is measured. After measuring them in the supernatant, the supernatant was discarded, and the pellet of BIND-PETase was resuspended for reuse.

Test

We encountered failures in collecting the enzyme and in accurately measuring the enzyme activity after collecting when we first undertook this experiment.

First, the pellets of BIND-PETases appeared to flow out when the supernatant was removed due to excessive force. Second, the amount of BIND-PETase used in this assay was too small to measure accurately. Additionally, the recovered BIND-PETase pellet contained pNP, the reaction product, surrounding it, which was visually confirmed by the fluorescent yellow color of the pNP.

Learn

We identified several areas for improvement in the method. First, the pNPB Assay utilizes a small volume of BIND-PETase suspension (10 µL), making the impact of BIND-PETase loss during recovery significant. Additionally, conducting the second and subsequent reactions with pNPs surrounding the collected BIND-PETase pellet further complicated the accurate measurement of Reusability.

≪cycle 3-2-2: Reusability Assay≫

Design

We aim to conduct a more accurate Reusability Assay. To achieve this, it is essential to minimize the effects of BIND-PETase loss during recovery and the residual pNP, a degradation product of pNPB, as noted in cycle 3-2-1.

Build

Therefore, we increased the volume of the BIND-PETase suspension used in the reaction to minimize the impact of pellet loss when the supernatant was removed. We also measured the optical density (OD) of the recovered BIND-PETase pellets after resuspending them in PBS, evaluating the activity per OD to assess the change in activity upon reusing BIND-PETase.

Additionally, the recollected BIND-PETase pellet was washed with PBS to remove as much of the residual pNP as possible as we do, thereby avoiding contamination in the next reaction cycle. For detailed protocols, refer to Experiments.

Test

We successfully measured the activity after reuse more accurately than in the previous cycle.
The increase or decrease in activity at the time of use varied among the different variants (refer to Results).

Fig. 4.1.24. Reusability of BIND-PETase variants

Learn

It was experimentally demonstrated that all variants of BIND-PETase possess reusability, which is important for implementation as well as storage activity, and was a crucial factor in the selection of BIND-bearPETase. Those with decreased activity upon reuse are thought to have lost activity due to structural changes caused by either being used in a reaction once or by the recovery process through centrifugation.

≪cycle 3-3-1: Plastic Pellet Degradation Assay≫

Design

In the implementation of BIND-PETase, understanding its degradation activity on plastic pellets is crucial. Therefore, we investigated the degradation activity of PET pellets using BIND-PETases with high storage activity and reusability (ID 23, 24).

The plastic pellets were provided by esa Inc., a plastic recycling company. For more detailed information on the interview with esa Inc., (refer to [IHP])(https://2024.igem.wiki/waseda-tokyo/human-practices/#section-1.-encountering-the-challenges-of-plastic-recycling). The pellets included both pure PET pellets and composite plastic pellets containing PP and PE in addition to PET, both of which were under consideration for our application.

Fig. 4.1.25. Pellets provided by esa Inc.

Build

BIND-PETase (ID23, 24) was suspended in solutions adjusted to pH 7 (PBS) and pH 9 (glycine-NaOH).

Test

The prepared pH 7 and pH 9 BIND-PETase suspensions were applied to the pellet fragments and left to stand at room temperature for 5 days (refer to Results).

After 5 days of reaction, the pellets were weighed and the percentage of change was determined.The results are shown in Fig. 4.5.5 and Fig. 4.5.6.

Fig. 4.1.26. Mass reduction rate of mono-plastic pellets(PET(N))

Fig. 4.1.27 Mass reduction rate of composite-plastic pellets(PETPEPP)

Learn

The degradation of mono-material and composite pellets was confirmed to some extent by BIND-PETase (ID23, 24). However, the pellets used in this experiment varied in weight, shape, and surface area, leading us to conclude that quantitative experiments are necessary to more accurately compare the activity of different variants.

≪cycle 3-3-2: PET Powder Degradation Assay≫

Design

We decided to conduct a more quantitative experiment than the PET pellet degradation experiment performed in cycle 3-3-1 to accurately evaluate and compare the PET degradation activity of BIND-PETase. For this, we planned to quantitatively analyze TPA, MHET, and BHET, the degradation products shown in Fig. 4.1.28, using HPLC.

We intended to use PET powder derived from PET bottles as a substrate. In this assay, we also tested BIND-duraPETase, the ancestor of BIND-PETase (ID23, 24), for comparison.

Fig. 4.1.28. Degradation Reaction of PET by PETase

Build

BIND-duraPETase, BIND-PETase (ID23, 24) were expressed to make suspensions in pH 7 (PBS) and pH 9 (glycine NaOH). PET powder was made by sanding a PET bottle as shown in Fig. 4.1.29. The PET powder is shown in Fig. 4.1.30.

Fig. 4.1.29. Making PET Powder from Plastic Bottles

Fig. 4.1.30. PET Powder

Test

PET powder was reacted with BIND-duraPETase and BIND-PETase (ID23, 24) at pH 7 or 9 for 3 days.The results are shown in Fig. 4.1.31 and Fig. 4.1.32. (refer to Results).

Fig. 4.1.31. PET Powder Degradation Degree(pH7)

Fig. 4.1.32. PET Powder Degradation Degree(pH9)

Learn

By measuring the amount of PET degradation products using HPLC, we were able to quantitatively verify the PET degradation activity of each BIND-PETase. The results allowed a more accurate comparison of the PET-degrading activity of each mutant, which was very important in determining the BIND-bearPETase.

BIND-bearPETase was determined from the BIND-PETase designed through the DBTL cycle by performing multiple assays. See Result for the detailed process of BIND-bearPETase selection.

References

Yinglu C. et al. (2021), Computational redesign of a PETase for plastic biodegradation by the GRAPE strategy, ACS Catal, 1340–1350. https://doi.org/10.1101/787069 ↩︎ ↩︎
Zhu B. et al. (2022), Enzymatic Degradation of Polyethylene Terephthalate Plastics by Bacterial Curli Display PETase, Environ. Sci. Technol. Lett. 9(7), 650-657, https://doi.org/10.1021/acs.estlett.2c00332 ↩︎ ↩︎
Ahan R. et al. (2019), Cellular Biocatalysts Using Synthetic Genetic Circuits for Prolonged and Durable Enzymatic Activity. Chembiochem.20(14), 1799-1809,
https://doi.org/10.1002/cbic.201800767. ↩︎
Yunpu J. et al. (2022), Hydrophobic cell surface display system of PETase as a sustainable biocatalyst for PET degradation, Microbiotechnology,Vol.13, https://doi.org/10.3389/fmicb.2022.1005480 ↩︎
Thomas H. et al. (2024), Simulating 500 million years of evolution with a language model, bioRxiv, https://doi.org/10.1101/2024.07.01.600583 ↩︎
Miguel G. et al. (2022), DOCKSTRING: Easy Molecular Docking Yields Better Benchmarks for Ligand Design, J. Chem. Inf. Model. 62, 15, 3486–3502, https://doi.org/10.1021/acs.jcim.1c01334 ↩︎
Smith S. et al. (2020),Assessing multiple score functions in Rosetta for drug discovery, PLoS One.15(10): e0240450.https://doi.org/10.1371/journal.pone.0240450 ↩︎
Oleg T. et al. (2010), AutoDock Vina: Improving the speed and accuracy of docking with a new scoring function, efficient optimization, and multithreading, Computational Chemistry, Vol. 31, Issue 2, 455-461, https://doi.org/10.1002/jcc.21334 ↩︎
Patrick E. et al. (2023), Plug & Play Directed Evolution of Proteins with Gradient-based Discrete MCMC, Mach. Learn.: Sci. Technol. 4 025014, https://doi.org/10.1088/2632-2153/accacd ↩︎
Zeming L. et al. (2023), Evolutionary-scale prediction of atomic-level protein structure with a language model, Science, Vol. 379, Issue 6637, 1123-1130, https://www.science.org/doi/10.1126/science.ade2574 ↩︎
Thomas H. et al. (2019), The EVcouplings Python framework for coevolutionary sequence analysis, Bioinformatics, Vol. 35, Issue 9, 1582–1584, https://doi.org/10.1093/bioinformatics/bty862 ↩︎
Shi L. et al. (2023), Complete Depolymerization of PET Wastes by an Evolved PET Hydrolase from Directed Evolution, Angewandte Chemie International Edition, 62(14), https://doi.org/10.1002/anie.202218390 ↩︎
John J. et al. (2021), Highly accurate protein structure prediction with AlphaFold, Nature, Vol. 596, 583–589, https://doi.org/10.1038/s41586-021-03819-2 ↩︎
Sulaiman S. et al.(2012), Isolation of a Novel Cutinase Homolog with Polyethylene Terephthalate-Degrading Activity from Leaf-Branch Compost by Using a Metagenomic Approach, Appl Environ Microbiol., 78(5), 1556–1562, https://doi.org/10.1128/AEM.06725-11 ↩︎
The PyMOL Molecular Graphics System, Version 3.0.3 Schrödinger, LLC, http://www.pymol.org/pymol ↩︎