Figure 1. DBTL circle

In project design and experimentation, we continually overcome the various challenges we encountered in order to prioritize safety, effectiveness, and the user satisfaction. By employing engineering approach, we quantify these challenges and define our target solutions, conducting design cycles at both wetlab and drylab levels. There were limited options available to us in terms of how to engineer the probiotic to target the pathogenic Fusobacterium nucleatum. Ultimately, whether we succeed or fail, we remain committed to the concept of “creating to learn, creating to apply” as we continuously iterate and optimize. This not only guides us in solving our own problems but also enables other teams to gain valuable insights from our work.

Figure 2. How do we target the bacteria?

How do others handle the similar problem?
When it comes to using bacteria to target another bacteria, in other words, achieving cell-cell adhesion, various cell-cell adhesion toolboxes have been developed over the years. In 2015, Pin˜ero-Lambea et al. [1] constructed a novel method of cell-cell adhesion using Synthetic Adhesins (SAs), which guided bacteria to adhere strongly, rapidly, and specifically to the target antigen of cells. In the same year, Todhunter et al. [2], O'Brien et al. [3], and Koo et al. [4] achieved cell-cell attachment through chemically modified approaches, but their methods lacked genetic basis and were difficult to pass on between generations of cells. In 2016, Cachat et al. used a calcium-binding protein-mediated adhesion, but calcium-binding proteins might interact with other substances in the body, lacking specificity [5]. Also in 2016, an iGEM team LMU-TUM_Munich developed a cell glue that enabled 3D printing of cells through biotin-avidin coupling. Undoubtedly, this was a fantastic idea, but it still could not be used for our therapeutic purposes.

Figure 3 Different ways to achieve cell-cell adhesion. a. 2016 LMU-Minich. b. 2018 Glass,D.S. et al c. our solusion.

Initially, our favored approach was to achieve cell-cell adhesion using antigen-antibody pairs. Antibodies have excellent binding affinity and specificity to antigens. Therefore, if we wanted to target bacteria at bacteria, using antigen-antibody binding seemed to be a promising strategy. However, considering the complex structure of antibodies and the challenges of proper folding in prokaryotic expression systems, using Escherichia coli to display nanobodies became a secondary option. The E. coli expression system is the most commonly used system for expressing nanobodies and offers both periplasmic and cytoplasmic expression methods, each with its own advantages and disadvantages. Additionally, some engineered oxidative bacteria strains have been developed, but they require further experimental support. In 2018, Glass et al. [6] developed a cell-cell adhesion toolbox that utilized antigen-antibody binding. However, their goal was to enable adhesion between two engineered bacteria to form multicellular systems, and they used fixed antigens and antibodies. What we are aiming for, on the other hand, is to target the wild-type pathogenic bacteria, which is not compatible with their toolbox. While we can draw inspiration from past surface display techniques to achieve cell-cell adhesion, the development of specific adhesion factors requires our search, screening, and validating.

Design Circle 1

Screening of bacteria-targeting fragment
In order to target Fusobacterium nucleatum (Fn) with engineered bacteria, we would like to realize the bacterium(engineered bacteria)-bacterium(Fn) targeting through protein-protein interactions on cell surface. Therefore, screening and designing more appropriate donor-acceptor protein combinations are important for the intensity and specificity of targeting. To achieve this goal, we propose two different adhesion methods:

1.Utilize the principle of Fn colonization of CRC. Express the special receptor of CRC on the surface of engineered bacteria to achieve the adhesion between engineered bacteria and Fn.
2. Utilize the principle of self-assembly of Fn pilus. Express monomers of Fn pilus on the surface of engineered bacteria, so that they participate in the self-assembly of Fn pilus, thus realizing the combination of two bacteria.

Figure 4. Two methods to target Fn

For method 1, we knew that Fn adhered to CRC by the binding of FadA (Fn pilus) and E-cadherin (on the surface of CRC), to make it brief, we named it “E-cad – FadA principle”. Thus, we expected to display the key fragment of E-cadherin on the surface of the engineered bacteria, so that the engineered bacteria could accomplish the targeting mission by E-cad – FadA principle. However, considering the large size and complexity of the E-cadherin, we planned to truncate it. Mara Roxana Rubinstein et al [1] have found the key structure for E-cadherin binding to FadA, which is located in EC5. And Pep-11, with 11AA (ASANWTIQYND), is the shortest intercalating peptides for binding FadA. Therefore, EC5 and Pep-11 is our two candidates for Fn targeting.

As for method 2, mFadA, the monomer of FadA, can self-assemble to form long-chained pilus through the head-to-tail connection[2]. If we expressed the mFadA monomer on the surface of the engineered bacteria, it would participate in the self-assembly of the Fn pilus, thus attaching the engineered bacteria onto the surface of Fn as well, and realizing the tight junction between bacteria. In addition, since only a small structure of mFadA is related to self-assembly function, we discarded the irrelevant domain and obtained a short peptide of 27AA called B-domain, which has better functional specificity.

Figure 5. Four candidates to target Fn

To evaluate the above four candidates, we used Rosetta to dock each of them with wild-type mFadA.

Note: The FadA sequences of different species of Fn are slightly different. For safety reasons, we used Fn ATCC10953 to replace dangerous ones and the sequence of mFadA also came from this subspecies [1]. The spatial structure was subsequently predicted with the use of Alphafold2.

For EC5:Rosetta local_docking,nstruct=10000, repeated 3 times.
For Pep11:Rosie online server[5], RosettaDock-5.0,nstruct=10000, repeated 3 times.
For mFadA:Rosetta local_docking,nstruct=200000, no repeat.
For B-domain:Rosetta local_docking,nstruct=10000,repeated 3 times

In addition, we have conducted detailed experiments on the autonomous assembly of bacterial hairs. According to the experimental evidence, FadA can be self-assembled after 5 min.

Figure 6. Docking results of four candidates.

The docking results are shown in the following pictures:

Figure 7. Docking results of four candidates.| a: Diagram of the binding modes of the four peptides. b: Adhesion factors based on self-assembly principle have higher binding energies ( The data were analyzed using student’s T-test; *: P<0.05; **: P<0.01; ***: P<0.001).

In addition, we have done extensive work on computational structural biology.We characterized the above four proteins based on the GROMACS 54A8 force field, and verified the results of the above molecular docking by analyzing the key parameters of the above four complexes, such as RMSD, RMSF, Rg, and hydrogen bonding, and analyzing the binding free energies using mm-GBSA.

Figure 8. Molecular dynamics analysis results of EC5.| a. RMSD of EC5 protein complexes; b. Combined Rg and RMSD analysis of EC5 protein complexes. c. RMSF of the EC5 protein complex. d. Analysis of protein slewing in the space of EC5 protein complexes.e~f. Hydrogen bonding between the EC5 protein complex and the solvent and internal hydrogen bonding. g. Covariance matrix of EC5 protein complexes. h. Raschel diagram of the EC5 protein complex.( The dramatic fluctuations in the figure are caused by the cyclic boundary of the water box.)

It can be seen from the docking results that the adhesion modes based on principle ①, have lower binding energies, while the other two based on self-assembly principle demonstrate higher binding energies. It suggested that adhesion factors based on self-assembly principle have a better ability to bind to bacterial pilus. Therefore, it guided us to consider the display of bacterial pilus monomer or fragment structures on the surface as the basic method for targeting Fn.

Design Circle 2

mFadA Truncating & Rational Design
After determining the self-assembly method to target Fn, our program design became clearer. Through the surface display technology, the mFadA were fused with the surface membrane proteins of the engineered bacteria, so that the engineered bacteria could participate in the pilus self-assembly of Fn, and achieve bacterium-bacterium targeting.

Figure 9. Utilizing the self-assembly of pilin monomers for bacterial-bacterial targeting.

However, previous work on the screening of targeting fragments woke us up to the fact that mere surface display of mFadA monomer was ill-considered. The A-domain is used to target cancer cells, the B-domain is used to bind FadA, but mFadA monomer has both A and B fragments, it can also assemble itself [1]. Therefore, if we just display mFadA monomers on the surface of engineered bacteria, mutual adhesion between engineered bacteria may occur, forming bacterial clumps and precipitating, which is an unwanted result.

Figure 10. Big problem of displaying mFadA directly.

To prevent the self-assembly of the displayed mFadA, we need to delete the unrelated structural domains and remain only the functional structural domains that bind to the FadA.

The structure of mFadA is shown below. It is formed by two inversely parallel α-helices. These mFadA are connected head- to- tail to form a long chain, and about 30 AA near the N-terminal plays an important role in self-assembly. However, exactly where to make the truncation remains a problem that needs to be solved.

Figure11. mFadA can self-assemble together head-to-tail into long fibers.

We tried to truncate from the N-terminal position 20 to 30 and obtained 10 mFadA truncated fragments, which were referred to as nominated “B-domains”, and docked with mFadA monomers to verify the ability of different B-domains to target mFadA. The Evaluation index was interface_score(I_sc)

Figure 12. Truncate mFadA at different positions to divide into two function domains.

Utilize Rosetta to perform local_docking with mFadA for each of the 10 B-domains, nstruct=10000 (for each), take the value that results in the smallest I_sc and repeat 3 times.

What we learn & Learn from us
As we expected, the longer the retention length is, the higher the binding energy is(Fig. 5a). Among them cut28 and cut29 have higher binding energy. However, it can be seen that the largest contribution to the score is Arg27(Figure5b), which is not encapsulated in the centre of the active site and will not contribute to the self-assembly of the pilus. More importantly, we looked at the active site centre microinteractions (Figure5d), and the main hydrophobic interactions and secondary hydrophobic interactions play a major contribution to the pilus self-assembly. The leucine chain consisting of L7, L11, L14, L21, L53, L76, L84 plays the main hydrophobic binding role, and the salt-bonded shell consisting of ionic bond interactions wraps the hydrophobic centre to form a strong and stable bond. Figure5c is the microscopic interactions of mFadA reported in the literature[3], which is consistent with our conclusion.

Utilizing dry experiments to guide the design of wet experiments is common in a variety of synthetic biology work. Like our work above, dry experiments are an excellent aid. Science is not arbitrary or idealistic; it pursues a series of rigorous logical chains. We have taken this into account and have carefully explored where the best place to cut is, rather than making arbitrary decisions and conducting wet experiments. We hope you can glean some insights from it.

Figure 13. Analysis of docking results with different B-domains. a, the binding energy of different B-domains and mFadA. b, the arginine has no contacts with other residues. c, the micro contacts reported in article. d, microscopic interactions derived from dry experimental results.

1.The residue numbering is based on the complete mFadA sequence, rather than renumbering after truncation of domain A or B.
2.The protein sequence of mFadA can be referred to in this literature[2].
3.The information regarding interactions related to self-assembly mentioned above comes from literature[3].

Design Circle 3

Design and Structural Optimization of Fishing Rod Protein.
This is the “fishing rod protein” we designed to “fish” for specific bacterial species like nucleotide shuttle rods. It consists of a signal peptide (white), membrane protein (pink), linker peptide (cyan), and bacterial pilus fragment (yellow). Below is the process of how we chose and optimized the structure of the fishing rod protein.

Figure 14. The structure of fishing rod protein

Due to the oriented characteristics of bacterial pilus self-assembly, it would have been ideal to find a membrane protein for N-terminal surface display on BL, as this would have minimized spatial hindrance.

Figure 15. Function of bacteria – targeting fishing rod protein

However, we were unsuccessful in finding such a protein. Instead, we could only identify a membrane protein, GL-BP, for surface display on BL that allowed passenger protein insertion at the C-terminus. This created significant spatial hindrance for our bacterial pilus self-assembly, requiring our pilus to reverse direction by 180° to bind with the FadA pilus on the surface of the target nucleotide shuttle rods.

Figure 16. Our solution to overcome the challenge of spatial hindrance

To address this issue in the design of our fishing rod protein, we incorporated a lengthy hydrophilic linker to ensure sufficient flexibility to induce the pilus fragment to bend 180°. The vectorial nature of pilus assembly is of utmost importance and can often be overlooked. If you find it challenging to comprehend, we will provide a more detailed explanation in our upcoming hemolysin design.

Figure 17. The structure of GL-BP

Based on our investigation, both iGEM17_TJU_China and iGEM22_Shanghai_HS have utilized the GL-BP protein as the membrane protein for surface display on BL. Its associated ABC protein is an important membrane protein that actively transports specific substances across the cell membrane using energy in the form of ATP. Various ABC proteins are present on the cell membrane. GL-BP is one of such ABC proteins that is widely expressed in bacteria belonging to the bifidobacteria, and these bacteria have the cellular functionality of expressing GL-BP on their surface[1].

We initially used the wild-type GLBP as the membrane protein (green) and designed the fishing rod protein by adding a (GSGSGSG)5 flexible hydrophilic linker (yellow) at the N-terminal end, along with the pilus fragment (orange). We hoped that this signal peptide-membrane protein-linker-pilus fragment “fishing rod protein” would participate in pilus self-assembly, therefore achieving “bacterium-bacterium targeting.”

We used AlphaFold2 to predict the structure of the fishing rod protein.

Figure 18. Structure prediction of Bacteria-Targeting Fishing Rod Protein

After multiple repetitions, we found that the B-domain is indeed embedded within a groove of the GL-BP protein. Typically, the structure at both ends of a fusion protein should avoid binding to prevent significant spatial hindrance or blocking of the intended activity, which could affect its proper functionality. Therefore, we were greatly concerned that such binding might lead to a decrease in the self-assembly activity of the pilus fragment. Consequently, we closely examined how the pilus fragment binds, hoping to replace some of the “culprit” residues. This part of the work will be presented in detail in the MODEL. The final results were not satisfactory.

Design Circle 4

Choosing a linker for fusion protein
With the advancement of molecular biology, numerous active proteins and peptides have been invented or discovered as drugs. Recombinant technology, known for its advantages of large-scale production and simplicity of operation, has found widespread application in various fields, including biopharmaceuticals. To obtain multi-target, multi-functional active proteins, it is necessary to fuse together two or more known functional proteins. This approach, which results in bi-functional or multi-functional fusion proteins, has become one of the new methods for drug development and biopharmaceutical research. It is particularly widely used in areas such as the preparation of bispecific single-chain antibodies (scFv) or antibody-conjugated drugs[1-5].

Figure 19. ADC drugs and PDC drugs

We initially used the wild-type GLBP as the membrane protein (green) and designed the fishing rod protein by adding a (GSGSGSG)5 flexible hydrophilic linker (yellow) at the N-terminal end, along with the pilus fragment (orange). We hoped that this signal peptide-membrane protein-linker-pilus fragment “fishing rod protein” would participate in pilus self-assembly, therefore achieving “bacterium-bacterium targeting.”

FK-13 exhibit a dual effect of antibacterial and anticancer properties. Even though it is an endogenously synthesized antimicrobial peptide, for safety and efficacy considerations, we aim to impart targeting capabilities to the FK-13 by introducing a targeting domain.

Figure 20. Foundation of fusion protein design

Firstly, we seek to understand the structure of the FK-13 and explore the possibility of constructing fusion proteins based on it. We do not want the introduction of linker peptides to disrupt the original structure of the FK-13. Therefore, we initially need to obtain the spatial structure of the original FK-13. We will then add different linker peptides at both the N-terminus and C-terminus to investigate the structural conservation of the original FK-13.

Figure 21. To provide a positive charge center for FK-13.

Secondly, we have also observed that altering the interaction of antimicrobial peptides with bacterial cell membranes plays a crucial role in modulating their antimicrobial activity. Yang Yanli et al. [6] replaced Glu16, Asp26, and Glu36 with Gln16, Asn26, and Gln36, respectively, increasing the net positive charge from +5.8 to +9.0. This modification raised the positive charge level of LL-37 without altering its spatial structure. Ga-gnon et al. [7] found that peptide chain length is also related to antimicrobial activity. For peptides of the same length, longer peptides with more positive charges exhibit better antimicrobial effects. Therefore, we intend to eliminate the linker at both ends of the killing peptide. The introduction of negative charges near the killing peptide should not reduce its antimicrobial activity. To ensure antimicrobial activity while designing fusion proteins, we have designed two types of linkers where positively charged lysine residues flank the killing peptide.

Therefore, we constructed four different sequences: LinkerG as a flexible linker and LinkerA as a rigid linker.

Figure 22. Sequences of FK-13 fusion protein design feasibility tests.

To maximize the preservation of TP's cytotoxic activity, we modified the sequence of LinkerA, resulting in LinkerB, thereby enhancing TP's positive charge and achieving stronger cytotoxicity.

We conducted several rounds of antibacterial and anticancer experiments using purified preparations of modified FK-13. After completing with the experiments, we focused on analyzing the cell death induced by FK-13 using CCK-8, flow cytometry and other analytical tools, and successfully calculated the IC50 of TTP against SW480 = 124.034 μg/mL.In addition, FK-13, as a central part of this project, it also played a role in killing Clostridium nucleatum. We also characterized its antimicrobial activity, and according to the results in the figure below, FK-13 has a strong antimicrobial ability compared to the 4 common antibiotics. This also proves that our rational modification is reasonable and feasible, which also lays an important foundation for our next work.

Figure 23. Results of TTP anti-tumor cell and antibiosis assay.| a.Detect the apoptosis of SW480 cells treated with FK-13 by Flow Cytometry. b~d. IC50 of FK-13 assay for SW480. e. Antimicrobial activity assay of FK-13.

The structures of the four sequences were predicted using AlphaFold2, and the results are as follows.

Figure 24. Structure predictions of FK-13 with different linkers.| a: High rigidity of FK-13; b: The flexible linker peptide still maintained flexibility with FK-13; c~d: E E would impair the anticancer activity of FK-13, and reversing the linker peptide would avoid this effect.

It can be observed that the LL-37 truncated peptide (FK-13) possesses a conserved α-helical spatial structure. Whether flexible or rigid linker peptides are added at both ends, there is almost no impact on the original structure. Furthermore, the results also indicate that adding a flexible linker results in lower confidence at both ends of the peptide chain, indicating higher flexibility. This highlights the need to consider potential interference between different functional segments when designing fusion proteins.

Based on the above results, we identified that FK-13 exhibits a simple α-helical structure, which is structurally conserved and less susceptible to interference, making it a potential candidate for designing fusion proteins. Additionally, when using a rigid linker, the α-helical structures of both can seamlessly combine, reducing uncertainty in fusion protein design. Furthermore, we made attempts at both the N-terminus and C-terminus, paving the way for our final dual-target peptide design, serving as the cornerstone for fusion protein design work.

Design Circle 5

Design of BTP and TTP
The construction of recombinant fusion proteins involves two factors: the constituent proteins and linker peptides. The selection of constituent proteins is relatively straightforward and is based on the desired functional properties of the fusion protein product. However, choosing the appropriate linker peptide can be challenging and is often overlooked in the process of designing fusion proteins. Failure to use linker peptides between functional segments can lead to adverse outcomes, such as misfolding of the fusion protein, low yields, or compromised activity. Therefore, the rational design and selection of linker peptides are crucial for the construction of fusion proteins.

Linker peptides are typically categorized as flexible linker peptides and rigid linker peptides, each with its own advantages and disadvantages. Flexible linker peptides like (GGGGS)n offer higher flexibility. However, the soft nature of flexible linker peptides allows the functional proteins at both ends to move freely. When adopting a side-by-side conformation, this "shoulder-to-shoulder" structure results in compact coiling of the fusion protein, making it prone to dimer formation. The close-packed conformation can also lead to the entwining or shielding of active sites, potentially causing a decrease in activity.

On the other hand, rigid linker peptides effectively control the relative positions and distances between the functional proteins at both ends, adequately separating different structural domains and minimizing their mutual interference. The most commonly used rigid linker peptide is the (EAAAK)n helical sequence, which forms an α-helix secondary structure that is less prone to bending, ensuring relative stability of the functional protein spacing.

However, Wu et al. [1] discovered self-cleavage properties of rigid linker peptides under specific conditions during their study of chitin fusion. Breakage occurs at pH 6 to 7, attributed to the arrangement of amino acids like EAAAK, which can form a stable hydrophilic α-helix structure. The key forces involved in this process are the Glu-...Lys+ salt bridges, with hydrogen bonding being a crucial force in forming these salt bridges. Therefore, pH significantly affects the stability of salt bridges, with the weakest interactions occurring in neutral solutions, making them prone to breakage, while high salt environments help stabilize the salt bridges, preventing cleavage [2].

In order to achieve targeted killing of Fn (Fusobacterium nucleatum) and CRC (colorectal cancer), we aim to design a targeted fusion protein based on the LL-37 truncated peptides. We have chosen the B-domain of mFadA as the structural unit for targeting Fn (the reasons are detailed in #Targeting Peptide Screening#), and for targeting CRC, we have selected HlpA (the reasons are detailed in #Q&A#). Now, our challenge is to figure out how to construct a fusion peptide that can effectively target bacteria.

Figure 25. The notices considered in TTP and BTP design. a, the self-assembly has vectorial property, B-domain would better to be placed at the N-terminal of the designed fusion proteins. b, the cytotoxic domain and the targeting domain should ideally be oriented in different directions.

We constructed BTP (Bacterial Targeting Peptide) using both flexible and rigid linker peptides with the aim of ensuring that the linker separates the targeting domain and the cytotoxic domain, allowing them to perform their respective roles without mutual interference. Since the self-assembly of the B-domain is directional, as shown in Figure 1a, to minimize spatial interference, it should be used as the N-terminus, with other functional domains added at its C-terminus to design the fusion protein, as illustrated in Figure 1b. For the purpose of "function domain isolation," the cytotoxic domain and the targeting domain should ideally be oriented in different directions.

In the design of TTP (Cancer Targeting Peptide) for targeting CRC, we also need to consider whether it should be added to the N-terminus or C-terminus. Additionally, the choice of linker is equally important. Therefore, this design circle will explore different approaches for designing BTP and TTP.

We constructed fusion BTP and TTP using both flexible and rigid linker peptides, and the sequences are as follows:

FK-13 exhibit a dual effect of antibacterial and anticancer properties. Even though it is an endogenously synthesized antimicrobial peptide, for safety and efficacy considerations, we aim to impart targeting capabilities to the FK-13 by introducing a targeting domain.

Figure 26. Sequence of HlpA, BTP and TTP. Notice: full sequence of HlpA can be found in registry: BBa_K4990007

First, we started with individual HlpA monomers and homodimers. Next, we used flexible linker peptides, forward and reverse rigid linker peptides, and designed fusion proteins both at the N-terminus and C-terminus of HlpA. Additionally, to ensure that the cytotoxic domain and the targeting domain are oriented in different directions, we designed an additional β-turn added to the N-terminus of HlpA, allowing the cytotoxic peptide to be adequately separated from the targeting structural domain.

We performed several rounds of anticancer experiments using purified and prepared TTP. Most notably, we analyzed TTP-induced cell death using CCK-8, flow cytometry, and other analytical tools, and successfully calculated the IC50 of TTP against SW480 = 136.03 μg/mL. the excellent IC50 somehow proved that the additional β-turn we added when designing TTP played a role!

Figure 27.Results of TTP anti-tumor cell assay.| Detect the apoptosis of SW480 cells treated with TTP by Flow Cytometry. b~d. IC50 of TTP assay for SW480. e. TTP is anti-tumor mainly through cell necrosis.

In addition to this, we validated the antimicrobial activity of TBP in the final stage of the project and compared it with DEH, which was no less potent than Amp against C. nucleatum.

Figure 28.Results of Antimicrobial Circle Experiment.

The structures of the nine sequences were predicted using AlphaFold2, and the results are as follows:

Figure 29 Results of structure prediction. HlpA is colored in cyan. Linker is colored in yellow. FK-13 is colored in red.

As shown in figure3,HlpA dimer have two helical segments from each monomeric subunit constitute an α-helical ‘body’ with two protruding β-ribbon ‘arms’ , which extend to bind the heparin[3] and DNA helix[4]. 

What we learn & Learn from us
The linkers for Peptide ①, ②, and ③, as well as FK-13, are all located at the C-terminus of HlpA. It can be observed that although these fusion proteins possess some flexibility, FK-13 in most cases is oriented towards the β-ribbon 'arms.' We believe that this orientation may introduce spatial hindrance, thereby reducing the targeting ability. The ideal orientation for FK-13 is similar to that of Peptides ④ and ⑨, especially with the addition of the GPNG sequence in ⑨. These sequences exhibit flexibility and point in the opposite direction of the β-ribbon responsible for targeting, which aligns with our emphasis on "function domain isolation."

We recommend that future protein designers, when creating fusion proteins, always keep "function domain isolation" in mind. The methods we have employed include changing the types and lengths of linkers, altering the order of fusion protein splicing, and modifying certain amino acids to achieve "function domain isolation." No one desires a protein with a functional surface that results in "1+1 > or < 2." Only when "1+1=2" is achieved, can it be considered the most controllable and predictable outcome. Synthetic biology is akin to building with blocks, and we hope that these functional modules can be stacked like building blocks, each performing its designated role, which has always been our ultimate goal.

Cytotoxic Peptide design circle 2

This is the third round of DBTL circle for the cytotoxic peptide design. Typically, when conducting a design circle to address the challenges encountered in engineering a solution in the "engineer" phase, we often go through a series of research steps to lay the groundwork for the work. However, this time is somewhat unique, as we have merged the previously designed BTP and TTP from the last round. Where did the inspiration come from, and why did we do this? Let's start with a discussion that led to this decision.

At that time, we had already completed the target peptide screening and had split the mFadA domain, obtaining the A-domain for targeting CRC and, more importantly, the B-domain for targeting Fn. Therefore, the BTP and TTP at that time were merely fusion proteins designed by simply joining the FK-13 with the A and B domains. However, one day , we discovered a flaw in the project design.

Figure 30. How we come up with the dual-targeting peptide

In simple terms, the dual-targeting peptide is assembled from the best BTP and TTP designed in the previous design circle. It consists of five parts: 1. B-domain for targeting Fn. 2. Linker A for separating functional domains. 3. FK-13 for killing Fn and CRC. 4. Linker B, which not only separates the functional domains but also ensures the positive center of FK-13. 5. HlpA monomer for targeting CRC. These segments have been individually uploaded as parts to the Registry, and you can find detailed descriptions of them in contribution.

Figure 31. Domains of DEH

Two DEH sequences were constructed sequentially and folded using Alphafold2. The best DEH was subjected to wet lab characterization.

FK-13 exhibit a dual effect of antibacterial and anticancer properties. Even though it is an endogenously synthesized antimicrobial peptide, for safety and efficacy considerations, we aim to impart targeting capabilities to the FK-13 by introducing a targeting domain.

Figure 32. Sequences of DEHs

Four sequences were subjected to structure prediction using Alphafold2, and the results are as follows,As shown in the figure, the rationally designed DEH has a better tertiary structure, which also endow it with stronger tumor targeting and killing ability.

The rationally evolved DEH solves the problem of dual-targeting functional crosstalk with better targeting. To prove that our DEH has a stronger ability to target cancer cells, we performed Pull down experiments on DEH using magnetic beads attached with heparin (heparin is an analog of HSPG glycoprotein), and the results are shown below. Since the mechanism of DEH targeting colorectal cancer cells is similar to that of TTR, we used the successfully characterized TTR as a reference for this experiment. According to the bands around 10KDa, it is known that there is a large amount of DEH binding, which indicates that our DBTL work is successful. And it was found that some DEH conditionally breaks to release FK-13 around 5KDa due to the PD condition of pH=6.7, which also proves the reliability of the conditionally broken linker peptide we chose before.

Figure 33. PD of DEH and TTR.

Based on the existing stem experiments, we performed prokaryotic expression of DEH and purified the DEH protein. Using this protein, we performed a large number of anti-tumor cell experiments, such as CCK-8 to detect cell activity, calculating the IC50 of DEH, and flow cytometry to detect apoptosis and necrosis of tumor cells. According to the above experiments, we measured the IC50 of DEH to be 174.43 μg/mL, which is slightly higher than that of LL-37, but much lower than that of 5-FU, the first-line drug for colorectal cancer in the current clinical treatment, which also indicates that our rationally-modified DEH has a good drug-forming property. According to the analysis of cell death by flow cytometry in the figure below, it can be seen that DEH induces tumor cell death mainly through the pathway of inducing necrosis of SW480 cells, which is also consistent with our design intention. The contents released from cell rupture could also further increase the level of immune infiltration of the tumor. Meanwhile, a large number of experiments have also proved that DEH is one of the best anti-cancer peptides designed by our team. These exciting results demonstrate that our rationally designed DEH has satisfactory anticancer properties.

Figure 34. Results of DEH anti-tumor cell assay.| a. Detect the apoptosis of SW480 cells treated with DEH by Flow Cytometry. b.Comparison of IC50 of rationally designed ACPs and two small molecule antitumor drugs against SW480. c. DEH is anti-tumor mainly through cell necrosis. d. Comparison of cell death in DEH, TTP, and LL-37 administration groups. e~f. IC50 of DEH assay for SW480.
Figure 35. Structure predictions of DEHs

It can be seen that if using the complete B-domain (①), there is an inability to form a complete α-helix. Therefore, the optimized DEH had the C-terminal five amino acids removed from the B-domain to ensure the stable formation of the α-helix structure at the N-terminus of DEH. As shown in ④, the added GPNG sequence once again played a role, targeting the Fn segment downward and the CRC targeting functional domain upward, effectively achieving "function domain isolation." Therefore, this optimized DEH was selected as our final targeting cytotoxic peptide.