Bispecific antibodies (BsAb) are a large family of molecules designed to recognize two different antigens. This property has been exploited to target two types of membrane receptors within a cell. Traditional bispecific antibodies are formed by combination of two half of different light-heavy chain antibody moieties (Fig. 1). In our project, we propose a novel bispecific antibody design that consists of two major components, a synthetic cyclic peptide and a native antibody (Fig. 1). The cyclic peptide is presumably conjugated to the antibody to form a secondary antigen binding site (bispecific). Since linear peptides are prone to proteolytic degradation and the conjugation efficiency varies depending on the circularization of target peptide, we employ a synthetic biology approach to design and build an intein splicing BioBrick that is capable of one-step cyclization and purification of the target peptide for subsequent conjugation (Fig. 2).

Figure 1Structures of bispecific antibody. A) Structure of conventional bispecific antibody. The primary and secondary antigen binding sites are made up of two half of the light-heavy chain antibody moieties. B) Structure of the novel design bispecific antibody. The primary antigen binding site is derived from the parental antibody, while the secondary antibody is derived from the conjugated synthetic cyclic peptide.
Figure 2Design and workflow of the novel bispecific antibody production. The novel bispecific antibody design consists of two major components, a synthetic cyclic peptide and a native antibody. The BioBrick for cyclic peptide is designed with two flanking inteins and a chitin binding domain (CBD) for purification. The intein spliced cyclic peptide is presumably conjugated to the Fragment crystallizable (Fc) region of the parental antibody, forming a secondary antigen binding site (bispecific). We employ a synthetic biology approach to design and build an intein splicing BioBrick that is capable of one-step cyclization and purification of the target peptide for subsequent antibody conjugation.

We are utterly aware that NSCLC is one of the common malignancies worldwide with a poor survival rate and it is the leading cause of cancer deaths in Hong Kong. Osimertinib, a third generation TKI that binds irreversibly to mutated EGFR, is currently employed as the first-line therapy to NSCLC patients with EGFR mutation. However, acquired drug resistance to treatment inevitably develops among patients, which limits the effectiveness of cancer treatment. Therefore, it is of great importance to develop an effective therapeutic strategy with the goal of increasing the survival of these patients. Currently, targeting Osimertinib resistance of NSCLC can be achieved by conventional bispecific antibodies which are now being tested in clinical trials. However, the production cost of the conventional bispecific antibody is high and the procedures tedious. Therefore, we employed a synthetic biology strategy to construct an innovative bispecific antibody called Polyneerab. In addition to designing and building synthetic BioBricks to express a functional cyclic peptide for effective antibody conjugation, we also characterized this novel bispecific antibody and examined its effects in a cell-based model as a proof-of-concept purpose. Our project harnessed the power of synthetic biology to develop a novel therapeutic bispecific antibody against resistant NSCLC, with the goal of overcoming drug resistance.

BioBrick Design

Phase 1

Our initial BioBrick design involves a linear EGFR-binding peptide (Ebp), CMYIEALDRYAF BioBrick combining the gene encoding the peptide and a His-tag for subsequent purification (Fig. 3). The choice of this Ebp is based on our molecular docking simulation presented in the Dry Lab page. The linear peptide is expected to be conjugated to the native monoclonal antibody.

Figure 3BioBrick design of pET-24d(+)-Ebp

Yet, the 2.4 kDa linear EGFR binding peptide was not successfully expressed, which might be due to the peptide insolubility, peptide degradation or other unknown factors.

To solve this problem, we searched for literature and adopted suggestions from advisers that cyclic peptide could reduce the possibility of peptide degradation and increase the stability of free peptides. Therefore, we included a cyclic form of CMYIEALDRYAF in our phase 2 BioBrick engineering.

Phase 2

To generate cyclic EGFR peptide in vitro, we design an Ssp-Mxe intein splicing BioBrick (Fig. 4). An intein is a segment of a protein that is able to excise itself and join the remaining portions with a peptide bond during protein splicing. The two split inteins located in the C-terminal and N-terminal of the Ebp will catalyze the excision of the inteins itself and ligate the C-terminal and N-terminal of the Ebp together to form a cyclic product. The BioBrick is also designed to be flanked by two chitin binding domains (CBD), which is used for purification. The use of CBD to substitute His-tag for purification is due to the less stringent elution condition required by CBD, which also facilitates the intein splicing activity. This design enables one-step purification and cyclization of the target peptide (Fig. 2). In this phase, we also included a scrambled peptide with random amino acids as a negative control for subsequent conjugation.native monoclonal antibody.

Figure 4BioBrick design of the Ssp-Mxe intein splicing BioBrick, pET-24d(+)-Ebp-Ssp-Mxe and pET-24d(+)-Scrb-Ssp-Mxe.

Our wet-lab progress has been hindered by the failed peptide purification. Although we had successfully expressed the Ssp-Mxe intein splicing BioBrick, we could not obtain the cyclic Ebp from the purification process. One possible cause of this outcome could be due to the fact that the Ssp-Mxe intein system is not suitable for generating short cyclic peptides. In response to this, we use another intein called Npu intein, which has been reported as an effective intein for short peptide cyclization.

Phase 3

In phase 3, we substitute the original Ssp-Mxe inteins with the Npu inteins, which is specialized for generating short cyclic peptides. This phase involves three BioBricks: the target Ebp, scrambled peptide, and a eGFP positive control. Similar to the Ssp-Mxe intein system, the BioBrick was flanked by two CBD for purification (Fig. 5).

Figure 5BioBrick design of the Ssp-Mxe intein splicing BioBrick, pET-24d(+)-Ebp-Npu-SsrA, pET-24d(+)-Scrb-Npu-SsrA and pET-24d(+)-eGFP-Npu-SsrA.

Owing to the time constraint, the purification and analysis of target peptide could not be finished within the competition period. Yet, the successful cyclization of eGFP control demonstrated that the Npu intein BioBrick is effective for cyclization of our target eGFP-binding peptide.

Engineering Cycle

We applied the engineering cycle in the design and construction of BioBricks. The engineering cycle involves three main stages: 1) Design and Build, 2) Test, and 3) Learn. During the competition period, we went through three complete cycles, incorporating our observations, experimental results, literature review, suggestions and insights from each iteration. Utilizing this iterative process, we successfully proved a one-step cyclization and purification system using eGFP.

Phase 1: Design and Build

Bio-synthetic linear EGFR-binding peptide as the solution

Based on the results of the molecular docking simulation conducted in the dry lab, we identified a robust peptide sequence, CMYIEALDRYAF, which demonstrated a strong binding affinity towards EGFR. With this in mind, we chose this peptide sequence in linear form as the EGFR binding peptide (Ebp) and incorporated it into our first BioBrick design.

Phase 1: Test

To test our BioBrick, we synthesized the BioBrick and carried it by an expression vector, pET24d(+). The expression vector was transformed to TOP10 competent cells for plasmid amplification and to BL21(DE3) competent cells for expression. However, we encountered challenges during the peptide expression stage, potentially stemming from issues, including insolubility during protein extraction, peptide degradation or other unidentified factors, leading to an unsuccessful expression of the linear Ebp peptide. To resolve this matter, further investigation and troubleshooting are required to pinpoint the root cause and devise an appropriate solution.

Figure 6Successful transformation of pET24d(+)-Ebp BioBrick to BL21(DE3) competent cells for peptide expression.

Phase 1 : Learn

The failure of the peptide expression led to our concerns about the peptide degradation due to their short sequence length. To solve these challenges , we conducted literature review on the stability of linear peptides and sought guidance from professionals in the field. It had become evident that linear peptides were susceptible to proteolytic degradation and exhibited inherent instability. Exopeptidases tend to cleave terminal amino acids, leading to their rapid degradation. Moreover, without an optimized conformation, linear peptides often exhibit a relatively low target binding affinity and reduced potency. Additionally, linear peptide structures tend to have lower lipophilicity, resulting in a poor cell permeability. These findings highlighted the challenges associated with utilizing linear peptide and led us to consider alternative strategies, such as peptide cyclization or incorporating structural modifications, to enhance the peptide stability, binding affinity, and cellular permeability in peptide-based bispecific design.

Phase 2: Design and Build

One-step purification & cyclization system

After a thorough literature review and consultation with experts, we decided to pivot our design towards cyclic peptides. Acknowledging the unexpected outcomes resulted from phase 1, particularly the problem of the peptide stability and proteolytic degradation, we took a bold step to develop a system that integrates both peptide purification and cyclization in just one step.

Our one-step purification and cyclization system are based on four major components:

The chitin column plays a crucial role in achieving one-step purification and cyclization. The split inteins are flanked by chitin-binding domains (CBD) that enable the affinity purification of the BioBrick. This CBD exhibits a strong affinity for chitin beads, facilitating the efficient and straightforward recovery of the cyclic Ebp from the crude cell extract.

Phase 2: Test

We performed a series of molecular cloning and protein expression experiments, starting from the restriction digestion of the pET-24d(+) expression vector and the synthesized BioBrick DNA fragments to yield compatible sticky ends for subsequent ligation, to transformation and IPTG induced protein expression. We confirmed the success of the restriction digestion through DNA gel electrophoresis. The restriction digested expression vectors and gene fragments were then purified and subsequently subjected to ligation where it incorporated desired BioBricks to the pET-24d(+) expression vector. The recombinant plasmids were then heat-shock transformed to TOP10 competent cells for plasmid amplification and BL21(DE3) competent cells for protein expression. The expression was performed by inducing BL21(DE3) cells with IPTG at 25°C overnight. The proteins were extracted by the lysis buffer and subjected to SDS-PAGE analysis. In this phase, we successfully expressed the CBD-intein-Ebp-intein-CBD fusion protein. However, when attempting the protein purification using the chitin column, regrettably, no peptide was detected from the elution. The possible reason for this outcome may be attributed to the unsuitability of the Ssp-Mxe inteins for generating short peptides.

Figure 7 SDS-PAGE Gel photo showing the protein expression of Ebp-Ssp-Mxe (Left) and Scrb-Ssp-Mxe (Right). The expected molecular weight (MW) of Ebp-Ssp-Mxe and Scrb-Ssp-Mxe is 54.5 kDa, which matches the band highlighted in red square. Lane 1, Protein MW Standards; lane 2, IPTG-induced soluble fraction; lane 3, IPTG-induced insoluble fraction.

Phase 2: Learn

We confirmed that Ebp was successfully expressed as the CBD-intein-Ebp-intein-CBD fusion protein as shown in the SDS-PAGE analysis. Notably, the intein BioBrick housing Ebp did not undergo degradation during protein extraction and was soluble to the lysis buffer. This observation supported the idea that with a stronger structural complexity, Ebp tended to be more stable. Yet, the short Ebp still posed a challenge for this intein BioBrick, as it might not be suitable for cyclizing such a short peptide.

Phase 3: Design and Build

Upgrade of the Ssp-Mxe inteins to Npu inteins

We hypothesized that the purification failure was linked to the unsuitability of the Ssp-Mxe inteins for cyclizing short peptides. To address this, we retained the intein system while opting for another superior intein, the Npu intein. We learned from literature review that the Npus intein is capable of expressing peptides as short as six amino acids, which made it a promising solution for us to continue the wet-lab. To verify the design, we additionally added a eGFP positive for this BioBrick. In this phase, we involved three BioBricks: the target and scrambled peptide, and a eGFP positive control.

Figure 8 BioBrick design of pET-24d(+)-Ebp-Npu-SsrA, pET-24d(+)-Scrb-Npu-SsrA and pET-24d(+)-eGFP-Npu-SsrA.

Phase 3: Test

The process of molecular cloning, transformation, protein expression and purification follow the same procedures as we had done in Phase 2. All groups including the Ebp, scrambled control and the eGFP controls were successfully expressed. By substituting the Ssp-Mxe inteins with the Npu inteins, we successfully obtained the cyclic eGFP from the Npu intein BioBricks. The calculated molecular mass of cyclic eGFP was 18 Da less than that of linear eGFP (26941 Da), resulting in a mass of 26923 Da, which indicated the linear eGFP has been modified into its cyclic form.

Figure 9 SDS-PAGE Gel photo showing the protein expression of Ebp-Npu-SsrA (Left) and Scrb-Npu-SsrA (Middle) and eGFP-Npu-SsrA (Right). The expected molecular weight (MW) of Ebp-Npu-SsrA and Scrb-Npu-SsrA is 34.6 kDa and eGFP-Npu-SsrA is 60 kDa, which match the band highlighted in red square. Lane 1, Protein MW Standards; lane 2, IPTG-induced soluble fraction; lane 3, IPTG-induced insoluble fraction.
Figure 10 LC-ESI-MS spectrum of elution fraction after chitin column purification showing the molecular mass of a cyclic eGFP (26921 Da).

Phase 3: Learn

In Phase 3, we successfully completed the purification process of the eGFP control group and were ready for purify the Ebp and scrambled control. However, due to time constraints, we were unable to finish the purification process of Ebp and Scrambled control. It is worth noting that, the cyclic Ebp generated by the Npu intein system had not yet been confirmed during the iGEM competition period, but we achieved a success in purifying and cyclizing the linear eGFP to cyclic eGFP. This outcome validates the capability of the Npu intein system to purify and cyclize the short peptide.

Since we cannot further proceed to the next round of the engineering cycle, in order to show the efficacy and robustness of our novel bispecific antibody, Polyneerab, we used a chemically synthesized cyclic Ebp for antibody conjugation to serve as proof-of-concept purposes. The bispecific antibody was subjected to functional assays, such as MTT cytotoxicity assay and enzyme-linked immunosorbent assay (ELISA). Details of the experiment results can be found in the Experiment page.

Figure 11ELISA binding assay of Polyneerab against (A) c-MET and (B) EGFR
Figure 12Combined treatment of Polyneerab with Osimertinib synergistically inhibited the growth of Osimertinib resistant HCC827 cells.

Based on this encouraging preliminary data, Polyneerab is a promising agent to overcome Osimertinib-resistance in NSCLC.