Team:NUDT CHINA/Poster

Poster: NUDT_CHINA

Predator Pro: a modularized toolbox for signal-controlled Targeted Protein Degradation

Poster presented by Team NUDT_CHINA 2020

Tianyi Zhang, Zhenyu Zhou, Changtai Xiao, Lunhao Ju, Xinlin Liu, Yongjiang Li, Huiying Liu, Qingyi Liu, Linjie Li, Ruoxi Wang, Yuxin Liu, Yingqian Ye, Junzi Gu, Haoyu Zhang, Yulan Chen, Zhiliang Pan, Hanxiao Feng, Guangyi Lin, Yanchen Gou, Yuxuan Wang, Weiqian Zhou.

Abstract

Exactitude temporal control of protein abundance is critical for the robustness and dynamics of synthetic circuits. While multiple approaches have been developed to manipulate the protein synthesis, few tools have been demonstrated to precisely control untagged protein degradation. Here, we present Predator Pro, a modularized and signal-controllable method for target protein degradation, on the basis of the Predator system we demonstrated in iGEM 2018-19. By rationally reengineer the Trim21 protein, we demonstrated that the interaction between Trim21 and antibody Fc domain can be replaced with other constitutive or inducible protein dimerization pairs. We demonstrated that constitutive DocS-Coh2 interaction or rapamycin-induced FRB-FKBP interaction enabled constitutive or drug-controlled degradation of untagged EGFP protein. As an effective expansion of the current synthetic biological tools for protein abundance control, this system may provide a modularized and convenient platform for controlled protein degradation, which might be applied in fundamental researches and clinical applications.

Introduction

● Spatial and temporal manipulation of protein abundance is the basis of designing and constructing complicated synthetic regulatory networks (Figure 1)

Figure 1. The mindset of mammalian cell synthetic biology

● Wide variety of synthetic biological tools have been developed to precisely control the protein synthesis on genomic, transcriptional, and translational level.

● Limited tools have been demonstrated to control protein degradation in synthetic biology (Figure 2).

Figure 2. Schematic representation of TPD methods in SynBio (A)Degron (B)PROTACs (C)Trim-Away

● Genetically encoded tool for signal-responsive degradation of untagged proteins is still needed.

Project Goals

This year, we oriented our efforts towards a highly modularized, genetically-encoded target protein degradation directly controlled by exogenous signals to manipulate untagged protein abundance.

Proposed Implementation

To sum up, PREDATOR Pro is tunable, reversible and highly modularized protein degradation tool with promising applications.

A: PREDATOR Pro would be helpful for SynBio researchers to manipulate untagged endogenous protein abundance in a freely selectable temporal and dose-dependent combinations.

B: For scientists working on gene function researches, PREDATOR Pro can also provide solution on developing cellular or animal model to knockdown the expression of specific gene in a tissue specific manner under given stimulus.

C: It's possible that our project might be used for therapeutic purposes. PREDATOR Pro can be used to control the target protein abundance within physical level, thus maintaining the original function of such protein.

Design

The Structure of Trim21

> RING domain: N-terminal, with E3 ligase activity;
> B-box domain, a coiled-coil dimerization domain;
> Coiled-coil domain;
> PRYSPRY domain, high affinity with constant Fc domain of the antibody.

The Mechanism of Trim21-mediated degradation

Once the antibody-antigen complex enters the cytosol, Trim21 binds the Fc domain of the antibodies and form a Trim21-antibody-antigen trimer, which mediate proteasomal degradation of the antigen.

Since it has demonstrated in previous research and our PREDATOR system that the interaction between the PRYSPRY domain and Fc is functionally independent to of other domains in Trim21, we modularized the PREDATOR system into two modules (a.k.a. Targeting module and Functional module). These two modules were connected by a protein-protein interaction-based interface.

We then hypothesized that:

the PRYSPRY domain of the Trim21 can also be replaced with other protein domains that can dimerize with other proteins under given signal input.

Phase I: Preliminary trial--Replaceability of the Trim21-antibody interface

Replaceability verification

Figure 1. Characterization of GFP PREDATOR Pro system based on the DocS-Coh2 interaction.

● Fig A. Schematic representation showing the design and function of GFP PrePro

● Fig B. Schematic representation showing the experimental workflow (left panel) and quantified GFP fluorescence heatmap 48 h post transfection of corresponding amount of plasmid (right panel).

● Fig C. Fluorescence images and quantified fluorescent intensity of HEK-293T cells 48 h post co-transfection with pEGFP plasmid and GFP PrePro/control plasmid.

● Fig D. Representative Western blotting determining the GFP plasmid protein abundance in HEK-293T cells 48 h post co-transfected with pEGFP plasmid and GFP PrePro/control plasmid (left panel), and quantified GFP protein level in three biological replicates (right panel). Relative protein level was calculated by normalizing the gray scale data of each group to the control group.

● Fig E. Representative fluorescence images in HEK-293T cell imaged 12-144 h post co-transfection with pEGFP plasmid and GFP PrePro /control plasmid.

These results suggested that the original PRYSPRY/IgG-Fc interface in the PREDATOR system could be replaced with other protein dimerization pairs while maintaining the protein targeting and degradation activity.

Phase I: Preliminary trial--Replaceability of the Trim21-antibody interface

New reporter system

Figure 1. Heatmap of fluorescence data detecting GFP and GFP-coupled firefly luciferase showing the degradation of GFP in RiPrePro-1 transfected groups with different concentrations of Rapamycin.

We used a dual luciferase reporter (pEFR), in which firefly luciferase (Fluc) was fused with GFP, to obtain more accurate GFP abundance changes (Figure 1A).

Fluc activity was used as an indicator of GFP abundance, and the activity of a constitutively expressed Renilla luciferase (Rluc) reporter was used to normalize the noise caused by irrelevant factors such as transfection efficiency, cellular protein synthesis and cell growth.

Results also showed significant lower (~50% lower) Fluc/Rluc ratio in GFP PrePro group comparing to the Ctr group (Figure 1B), indicating that GFP PrePro has impressive efficiency in degrading the target protein.

Phase II: Towards signal responsiveness

Increase of Rapamycin Concentration

Figure 1. Rapamycin-induced PREDATOR Pro system based on the FRB-FKBP interaction module.

To push our new PREDATOR Pro system towards signal responsiveness, we further engineered a Rapamycin-induced PREDATOR Pro system (RiPrePro-1) by replacing DocS-Coh2 with FRB-FKBP(Fig 1A).

To test whether the GFP levels can be tuned and continuously regulated by rapamycin, HEK-293T cells were co-transfected with GFP expressing plasmid and RiPrePro1.0 plasmid. Fluorescent imaging showed slight decrease (~20%) of GFP fluorescence in RiPrePro1.0 expressing group(Fig 1B and 1C) under 1 ng/μL rapamycin induction.

To improve the degradation effect of RiPrePro system, modeling group noticed that the degradation efficiency was highly sensitive to the parameter reflecting the interaction strength of the protein dimerization pairby by performing Sensitivity Analysis among all parameters in our model. In the case of RiPrePro system, such result can be improved by increasing the concentration of rapamycin or increasing the binding strength of the interface part of the protein.

Figure 2. Heatmap of fluorescence data detecting GFP and GFP-coupled firefly luciferase showing the degradation of GFP in RiPrePro-1 transfected groups with different concentrations of Rapamycin.

Therefore observed fluorescence image detecting GFP and GFP-coupled firefly luciferase showing the degradation of GFP in RiPrePro-1 transfected groups with different concentrations of Rapamycin.

As expected, the degradation effect could be significantly improved with the increasing concentration of rapamycin. However, huge amount of rapamycin would significantly reduce the cellular protein synthesis. Therefore, we come to another solution to multiplying FKBP copies for a higher intensity.

Phase II: Towards signal responsiveness

Two FKBP domains

With the implications obtained from the model, we changed the GFPnano-FKBP unit in the RiPrePro1.0 into GFPnano-FKBP-FKBP unit to further explored whether our Rapamycin-induced PREDATOR Pro system with two FKBP fragment (RiPrePro-2 or RiPrePro2.0) could show better performance.

Figure 1. Rapamycin-induced PREDATOR Pro system based on the FRB-FKBP-FKBP interaction module.

RiPrePro2.0 group is consists of HEK-293T cells, pEFR and plasmid expressing RiPrePro2.0.In Ctr group,there are HEK-293T cells, pEFR and empty vector.

Dual luciferase assay showed that the normalized GFP abundance in RiPrePro2.0 group was significantly lower than the RiPrePro1.0 group in most time points (Fig 1B), indicating an improved degradation efficiency under increased FKBP copies. Further analysis on RiPrePro2.0 further revealed that the GFP degradation activity of RiPrePro2.0 was dose dependent to the rapamycin concentration (Fig 1C).

Similarly, western blotting analysis showed ~40% lower GFP protein level in RiPrePro2.0 group (Fig 1D), which is in alignment with the dual luciferase assay results.

Moreover, 30%-60% reduction on Fluc/Rluc ratio in the RiPrePro2.0 group could be observed in different host cell lines we tested 48 h post 2 ng/μL rapamycin induction, suggesting good robustness of such RiPrePro system (Fig 1E).

In general, these results showed that our PREDATOR Pro system can be successfully engineered to degrade the target protein under the control of exogenous signals with decent performance and satisfying robustness.

The time-dependent characteristic of GFP PrePro system in silico and in vitro

In order to determine the experimental time of degradation of GFP PrePro and GFP, we simulated the function according to the experimental data, and drew the corresponding function curve. According to the above figure, we can see that the degradation rate of GFP is basically stable within 48 hours. In other words, the reaction between p18 and GFP was basically completed. Therefore, we think that the experimental group can take 48 hours as the node of experimental data collection, so as to obtain stable data and follow-up experimental observation results.

Interview to Prof. Xia & CCiC meetup

After the preliminary trial experiment of replacing PRYSPRY/Fc domain with DocS-Coh2 protein pairs, we consulted Professor Zanxian Xia and presented our data in CCiC meetup. We were suggested to change the GFP reporter into a Dual Luciferase Reporter system to normalize irrelevant factors affecting GFP abundance.

With this new reporter, the abovementioned experiments were repeated and the results clearly showed that GFP PrePro could degrade the target protein significantly.

Using our model to improve the degradation effect

Increase of Rapamycin concentration

To improve the degradation effect of our RiPrePro system, we approached our modeling group to figure out the most important factors affecting the PREDATOR Pro system. By performing Sensitivity Analysis (Figure A) among all parameters in our model, we noticed that the degradation efficiency was highly sensitive to the parameter reflecting the interaction strength of the protein dimerization pair(Figure B and C). In the case of RiPrePro system, such result implied that the degradation effect can be improved by increasing the concentration of rapamycin or increasing the binding strength of the interface part.

With the increment of rapamycin concentration, the degradation efficiency of RiPrePro did increase impressively. However, it has been well established that exorbitant amount of rapamycin would significantly reduce the cellular protein synthesis. Therefore, we come to another solution to multiply FKBP copies for higher binding strength.

Literature research to determine protein dimerization pairs regulated by exogenous signals

As a set of widely used and fully characterized heterodimerizing components, the dimerization of FK506 binding protein (FKBP) domain and the T2089L mutant of FKBP-rapamycin binding domain (FRB) could be initiated by external rapamycin signals. Therefore, we substituted the interface with FRB-FKBP protein interaction pairs.

Using our model to improve the degradation effect

Two FKBP domains

To promote the degradation efficiency, we turned to the model group for help. After performing Sensitivity Analysis of our system, they discovered that a parameter regarding the protein interaction intensity of FKBP and FRB was pivotal to the degradation efficiency. Therefore, we designed a Rapamycin-inducible PREDATOR Pro2.0 plasmid (RiPrePro2.0) composing GFPnano-FKBP*2 and HA-Trim21-FRB. It was observed that under 2 ng/μL rapamycin induction, the fluorescence intensity of HEK-293T cells co-transfected with pEGFP and RiPrePro2.0 plasmid was significantly lower than the control group transfected with pEGFP and empty vector.

Integrated HP

As a foundational advance project, our work in iGEM 2020 attempted to provide a novel toolbox to control the homeostasis of specific target proteins in mammalian cells. Hence, our lab work was mainly directed to serve synthetic biologists with new toys for synthetic circuit design, as well as scientists in other fields with new approach to develop cell models, etc. With these proposed applications in mind, our human practice work mainly focused on gathering ideas and suggestions from scientists and experts and collecting information on how our project could be shaped to meet their demands. Their supportive feedbacks helped us to reshape the project design and inspired us on the possible future application of our project as well.

Here we demonstrate how our integrated human practice interacts with our project design and the wet-lab experiments.

Model

Figure 1. Framework of the mathematical model

To understand the performance of our design comprehensively and offer guidance on wet-lab work, a mathematical model composed of two modules (protein interaction module and a protein degradation module) was constructed based on fundamental biochemical principles.

Figure 2. the side-by-side comparison of model prediction and experimental data

● The simulation showed that the model predictions matched well with the corresponding wet-lab data.

Figure 3. the influence of plasmid dosage and ratio on GFP degradation rate

● Model results showed that 1:1 is the optimal plasmid ratio and 0.25 mg dosage for both plasmids should be prioritized.

Figure 4. the sensitivity analysis of parameters in the model

● The Sensitivity Analysis of all parameters in the model calculate the time-dependent sensitivities (derivatives) of GFP with respect to each parameter.

k3f, the rate parameter characterizing the dimerization of FRB and FKBP, manifested the highest sensitivity across the board(Figure 3).

Results indicated that we could optimize the RiPrePro by increasing the value of k3f, or in other words, increasing the dosage of rapamycin or the number of FKBP domains. Considering the toxicity of rapamycin to cells, the future experiment focused on adding up the number of FKBP domains.

Interview to Prof. Xie

We approached Dr. Mingqi Xie after we obtained most of our data. Considering his record as a synthetic biologist, we were hoping that he could provide us some hint on how our project could be implemented in the real world. Through the interview, we first showed Dr.Xie the design of Predator Pro system, as well as the data we’ve already obtained. After knowing about our project, Dr. Xie praised our project for sufficient experiments and adequate data. Although some points still need to be improved, he thought that our project was logical and rigorous, and he had expectation that our work would be a useful tool for Synthetic Biology and Medicine. More importantly, he provided us some more suggestions on how to further improve our system. We were suggested to start looking for some therapeutic targets for our predator system. We were also suggested that further researches should be conducted on whether our system could be used to establish new disease cell or animal models..

Future Work

● New interface: Apply other signal-sensitive protein dimerization, Using CIBN-CRY2 interaction to achieve a blue-light-inducible PREDATOR Pro plasmid has been constructed and relevant experiments are within our plan.

● New target: protein-TAR DNA-binding protein-43(TDP-43), a hallmark of amyotrophic lateral sclerosis (ALS), to provide a promising solution to the treatment of ALS and other neurodegenerative diseases.

Partnership

Focusing on various ways for collaborations to occur, both online and offline

> A long-term cooperation mechanism with CSU_CHINA

Figure 1. The online and offline meeting with CSU_CHINA

> Offline meeting to deepen the communication between our teams. Both teams agreed on further and more concrete collaboration.

Figure 2. The partnership with CSU_CHINA

Partnership: CSU_CHINA helped us with our massive fluorescent images analyzing. With their brilliant R language skill, we gained a beautiful visualization of our fluorescent images (Fig 2A). As CSU_CHINA has limited access to the cadmium ion detection kits, we lent atomic absorption spectrometer and helped them measure the cadmium-uptaking levels of Synechocystis while in different growth states(Fig 2B).

Figure 3. The joint brochure for education

Furthermore, we designed our joint brochure for educational propaganda of synthetic biology (Fig 2).

Science Communication

● Spread the basic knowledge and mindsets of synthetic biology and iGEM to our college and Yali High School to deepen the public's understanding of synthetic biology

● Prepared some brochures about synthetic biology and our projects for students in lower age

● After we finished our project promotion video, we prepared a Chinese version and uploaded it online allowing others to know our team and our project

Parts

This year, we handed in 16 high-quality, well documented bio-brick parts, including all those we used in our project this year, and several brilliant designed others inspired by our iGEM projects within these two years.

Favorite basic parts

Our favorite Basic parts are Truncated Trim21 and CMV-Replaceable-1-Fluc-P2A-Rluc. Truncated Trim21 is the core of the newly registered Predator Pro system. This truncate protein maintained the E3 ubiquitin ligase activity, and provided an open interface for other protein dimerization pairs to be added.

Another part is a reporter system to quantify the abundance of specific target protein.

Favorite composite part

Our favorite composite part is Replaceable-1-Replaceable-2-P2A-TRIM21-Replaceable-3. It is a plasmid platform on which different targeting proteins and protein dimerization pairs can be easily installed to achieve the controllable degradation of specific target protein.

Part Collection

Our part collection “Predator” consists of 44 parts of our Predator Pro system and original Predator system we demonstrated in iGEM 2018-19. Using these parts, we can construct a Predator Pro system that can target any endogenous target protein regulated by exogenous signals.

Reference

1  Xie, M., Haellman, V. & Fussenegger, M. Synthetic biology—application-oriented cell engineering-2016-Current Opinion in Biotechnology. Current Opinion in Biotechnology 40, 139-148, doi:10.1016/j.copbio.2016.04.005 (2016).

2  Foss, S. et al. TRIM21-From Intracellular Immunity to Therapy. Front Immunol 10, 2049, doi:10.3389/fimmu.2019.02049 (2019).

3  Zeng, J., Santos, A., Mukadam, A. & Osswald, M. Substrate-induced clustering activates Trim-Away of pathogens and proteins. bioRxiv 225359 ,doi: 10.1101/2020.07.28.225359 (2020).

4  Barak, Y. et al. Matching fusion protein systems for affinity analysis of two interacting families of proteins: the cohesin-dockerin interaction. J Mol Recognit 18, 491-501, doi:10.1002/jmr.749 (2005).

5  Gaj, T. et al. ZFN, TALEN, and CRISPR/Cas-based methods for genome engineering. Trends Biotechnol 31, 397-405, doi:10.1016/j.tibtech.2013.04.004 (2013).

6  Bai, P. et al. A fully human transgene switch to regulate therapeutic protein production by cooling sensation. Nature Medicine 25, 1266-1273, doi:10.1038/s41591-019-0501-8 (2019).

7  Xue, S. et al. A Synthetic-Biology-Inspired Therapeutic Strategy for Targeting and Treating Hepatogenous Diabetes. Molecular Therapy the Journal of the American Society of Gene Therapy 25, 443, doi:10.1016/j.ymthe.2016.11.008 (2017).

8  Wang, H. et al. Treatment of chronic pain by designer cells controlled by spearmint aromatherapy. Nat Biomed Eng 2, 114-123, doi:10.1038/s41551-018-0192-3(2018).

9  Shao, J.et al. Smartphone-controlled optogenetically engineered cells enable semiautomatic glucose homeostasis in diabetic mice. Science Translational Medicine 9, eaal2298, doi:10.1126/scitranslmed.aal2298 (2017).

10  Kleiger, G. et al. Perilous journey: a tour of the ubiquitin-proteasome system. Trends Cell Bio 124, 352-359, doi:10.1016/j.tcb.2013.12.003 (2014).

11  Chassin, H.et al. A modular degron library for synthetic circuits in mammalian cells. Biotechnol J 10, 2013, doi:10.1038/s41467-019-09974-5 (2019).

Attribution

Model: Jie Cai, Yanchen Gou, Yongjiang Li, Zhiliang Pan, Haoyu Zhang. Our PI Xinyuan Qiu and Instructor Chuanyang Liu helped us with modeling ideas and methods.

Human Practice: Lunhao Ju, Zuyu Dai, Qingyi Liu, Xinlin Liu, Yuxin Liu, Linjie Li, Yulan Chen

Art Design: Guangyi Lin, Ruoxi Wang, Tianyi Zhang.

Wiki Design: Qinrui Jiang, Guangyi Lin, Ruoxi Wang.

General Support: The project was supported by the Department of Biology and Chemistry of College of liberal art and Science, National University of Defense Technology.

Project Support: PI Lingyun Zhu, Xinyuan Qiu; Instructor Chuanyang Liu, Lu Min, Lvyun Zhu, Xiaomin Wu, Jingyu Kuang, Wenying Li; Advisor Jiaxin Ma.

Fundraising help and advice: PI Lingyun Zhu, advisor Lu Min was in charge of expenses management.

Lab support: Department of Biology and Chemistry of College of Liberal arts and Sciences. Difficult technique support: Sisi Xie, Wenying Li, Xinyuan Qiu.

Human Practice & Wiki support: Our advisor Xinyuan Qiu, Chuanyang Liu provided us with solutions when we face difficulties in the process of wiki production.

Acknowledgement

Dr. Zanxian Xia

Dr. Mingqi Xie