C O N T R I B U T I O N

This year ,Team Peking make several contributions to future iGEM teams from all perspectives of Wet lab, Dry lab and HP.

Parts collection

Peking iGEM 2024 has designed and characterized a large number of basic parts and composite parts, and the relevant parts have been uploaded to parts.igem.org. If you are interested, please go to http://parts.igem.org/BBa_K53210##, where ## denotes the very numbers of each part.

Our Parts Collection mainly consists of three units: Aptamers, Proteases & Substrates and Output. Each unit has several subunits.

Beyond our project, these parts will be of reference and help to many iGEM teams who want to use GPA in situations that require de novo design of a molecule recognition and detection system.

Figure 1 Parts collection overview.

Verified thrombin aptamers

They are the aptamers that recognizes thrombin (which is already proved by early researches).

MAGA aptamers

They are the aptamers generated by our MAGA tools to bind thrombin.

Intact & split proteases and substrates

They are the intact proteases for our preliminary experiment, with codon oprimized for E.coli and 6x His tagged.

They are the split form of PPVp for preliminary experiments.

They are the substrate of PPV and TEV respectively. The linker sequence is designed for different cleavage events. There is also a PPVp substrate with MBP tag, so that we could purify the protein in the supernatant.

Proteases used in aptamer-split protease crosslinking

They are the split PPVp with a codon mutation(Phe → non-canonical amino acid azido phenylalanine), which will later be used in the chemical ligation with aptamers.

They are the split PPV—SA(strepavidin) fusion protein. As another protease—aptamer ligation method, SA is expected to bind biotin-modified aptamers with high affinity.

Output

This is the substrate as our output signal.

Design of cell-free system

The disease detection tool used in GPA is entirely cell-free. Compared to cellular systems (i.e., the detection and output proteins are expressed and become effective in live cells), the cell-free system offers the following advantages:

• Fast response and ease of use: Cellular detection systems require time for gene transcription and protein translation. In contrast, the cell-free system directly immobilizes the protein on the test paper, allowing for a faster response.

• Portability, stable storage and lower cost: Cell cultivation is an exhausting task, requiring stringent conditions for both growth and transportation. In contrast, the cell-free system is highly convenient for transport and can be stored for long periods without difficulty.

• Reduced risk of contamination: In cellular detection systems, contamination from environmental factors or microorganisms is always a concern due to the need for maintaining live cells. In contrast, cell-free systems eliminate this risk since there are no live cells involved.

• Scalability for mass production: Cell-free systems are easy to manufacture and scale up for widespread distribution, making them ideal for large-scale testing needs. In contrast, cellular systems require careful handling and controlled environments, limiting their scalability and increasing the complexity of mass production.

This kind of system showed outstanding efficiency, flexibility and portability. We hope future iGEM teams can be inspired by our design and think up more fascinating cell-free projects.

Insoluble protein purification

Proteins in most iGEM projects are usually soluble in water. Unfortunately, our split proteases are originally insoluble in water, and could be only detected in inclusion bodies. To handle this, our team investigated three methods to acquire water solution of the proteins, for reference of future iGEM teams:

• On-column renaturation in gravity chromatography: 8 M urea was used to denature the proteins, breaking the complex sturcture and making them soluble. Then the proteins were purified by gravity chromatography. After the proteins bound to the Nickel beads, a decreasing gradient concentration of urea was employed to renature the proteases. Through this way, the proteases were make soluble with considerable activity.

• Addition of soluble tags: The denaturation-renaturation operation caused activity loss after all. To improve protease activity, soluble tag MBP was added to the protease, which is realised by molecular cloning. After purification, a specific kind of protease was added to the protein solution to cut off MBP tag.

• Cell-free expression system: Another way of purification is to detour from cells. Alice® Cell-free Protein Expression System was adopted to express our proteins. This system allowed us to purify proteins without interferential proteins and other contents in living cells.

Molecular dynamics simulation

Based on the software GROMACS, we have systematically performed either single-element system (i.e., single cTEV or single nPPV) or compound-element system (i.e., both cTEV and nTEV exist to mimic the whole TEVp) molecular dynamics (MD) simulation for 10-ns under the OPLS-AA/L all-atom force field. By further trajectory analyses, we choose out proper Phe sites to be introduced in the split enzyme systems in order to carry out the following “clicking reaction”.

Differential equation modeling

For our own therapeutic system, we have established effective biochemical ODEs (mainly constructed from basic kinetic chemistry) and PDEs (mainly constructed from the Fokker-Plank diffusion model) based on some plausible assumptions, which presents insightful pictures for the mechanism of the censoring, amplifying, and diffusing process.

Following is the conceptual figure for our in silico experiments (Figure 2). Please refer to our Model page to see more details.

Figure 2 In silico Calculation overview.

1. Conceptual figure for Molecular Dynamics

2. Conceptual figure for Molecular Docking

3. Conceptual figure for ODEs system

AI for science

Since ChatGPT went viral in late 2022, a variety of generative models have emerged to assist human beings in several fields. With the progress of deep learning and machine learning technologies, AI systems are becoming increasingly capable of complex data analysis and predictions. For sure, there is no excption for biology. AlphaFold 3 has shown excellent performance on protein structure prediction and protein-ligand interaction prediction, reaching an accuracy of more than 90%. However, the GPA project needed a model that can give sequences of aptamers to given biomarkers, and the exsiting models did not meet our requirements. Therefore, our team developed our generative model: MAGA. It is the generation model named Make Aptamer Generally Applied (MAGA) that makes our model out-standing, which can predict UNA sequences with possibly high affinity for specific target ligand. For the results predicted by MAGA, we can then both validate them based on the molecular dynamics (MD) simulation in silico and the electrophoretical mobility shift assay (EMSA) in vitro.

MAGA is based on several previous studies: we refered to DTU-denmark 2023 iGEM project, leveraging the MAWS algorithm; we consulted a lot of literature, trying different generative methods and algorithms. Finally, we generalized the algorithm for aptamer generation and pushed our MAGA to github for future communications and applications. We sincerely hope future iGEM teams can make full use of it and make it a better tool.

Please refer to our Software page to see more details.

Human practices: a new paradigm

Our Integrated Human Practice (iHP) efforts have been instrumental in shaping and refining our project through continuous engagement with experts, stakeholders, and the broader scientific community. To ensure that our project evolves dynamically in response to real-world feedback, we devised the CIRR Cycle—Challenge, Investigation, Reflection and Refinement. This cycle became the core of our iterative process, guiding us through each phase of development and allowing us to integrate valuable insights into our work.

At the outset, our project focused on using biomarker detection for lung cancer diagnostics. However, through our first CIRR cycle, including an in-depth interview with Dr. Xiao Li from Peking University People’s Hospital, we identified the challenges of lung cancer’s heterogeneity. Dr. Li advised us to redirect towards nonsolid tumors, where biomarker detection has clearer advantages. Reflecting on her feedback, we refined our project and focused on chronic diseases, incorporating synthetic biology pathways for precise biomarker detection in in vitro diagnostics. This shift exemplified how our CIRR methodology allowed us to continually adapt and improve.

As we moved forward, further iterations of the CIRR cycle led us to optimize the technical aspects of our detection system. Our interview with Hangzhou Newblue Technology Co. helped us address challenges related to output methods. By reflecting on their feedback, we adopted aptamers as modular connectors, enabling faster screening and a more flexible design. This crucial refinement not only enhanced the practicality of our project but also maintained the efficiency of our colloidal gold kit output, which we had originally envisioned for home diagnostic use.

Finally, through our collaboration with Prof. Tao Liu and Assist. Prof. Liqin Zhang, we refined the linking mechanism for aptamers and split proteases, using thrombin as a proof-of-concept. Our final iteration focused on exploring market potential and drafting a business plan, informed by consultations with industry professionals. Each iteration of the CIRR cycle was like a river carving its way through the mountains—encountering obstacles yet always moving forward, eventually flowing into the vastness of real-world application. During this journey, we never ceased our dialogue with the world, and it was through this continuous interaction with the outside world that our project evolved, from a laboratory concept to a practical solution.