Overview:

To address bacterial speck of tomato caused by Pst, we have developed a cell-free rapid detection system. It targets two substances released by the bacteria: protease AvrRpt2 and phytotoxin COR. This system is based on a fundamental principle: Protease cleavage triggers rearrangement of the coiled-coils (CC) domain, bringing the split proteins attached to each CC domain into proximity, which in turn activates the split proteins.

Building upon this core mechanism, our design consists of the following parts:

(1) First input: protease AvrRpt2 (with a positive feedback loop to lower detection limit).

(2) Second input: phytotoxin COR secreted by Pst.

(3) AND gate design: the output occurs only when both inputs are present.

(4) Output: Three output modes are employed in our detection system. The first relies on luciferase, which emits luminescence upon the addition of its specific substrate. The second involves split spGFP, which emits fluorescence directly upon combination when exposed to its excitation light. The third integrates split trehalase, which converts trehalose to glucose for detection.

Basic Module:

Coiled-Coil (CC) domain is a ubiquitous structural motif in proteins. In previous studies, CC domains are extensively involved in protein dimerization and subunit assembly (Reinke et al., 2010).

Our project is designed based on antiparallel CC domains, which are characterized by a heptad repeat of amino acid residues, denoted as positions a to g in Fig.2. In antiparallel coiled coils, the key interactions that drive helix pairing are as follows: Firstly, residues at position g interact with g' residues, and residues at position e interact with e' residues through Coulombic interactions. These electrostatic forces play a crucial role in determining the specificity and orientation of the antiparallel helices. Secondly, the hydrophobic residues at positions a (a') and d (d') contribute to the stability of the CC domain. The a (a') residues always pack against d' (d) residues, resulting in a single type of hydrophobic layer. Together, these interactions involving electrostatic and hydrophobic forces enable the formation of stable antiparallel coiled-coil pairings (Oakley, 2001).

Taking advantage of this, we utilize the rearrangement of CC domains for enzyme assembly in our experimental design. The Target chain, connected to one part of a split protein, is initially linked to the Autoinhibitor chain with a protease-recognizable cleavable short peptide, while the Displacer chain links to the remaining part of this split protein. We engineered a reduction in the interaction stability between the Autoinhibitory chain and the Target chain, thereby facilitating the formation of a CC dimer between the Displacer chain and the Target chain.

When a protease cleaves the linker between the Target chain and the Autoinhibitor chain, the Target chain, which has a higher binding affinity with the Displacer chain, can therefore pair up with the Displacer chain. This displacement brings the split proteins connected to the Target chain and the Displacer chain close enough to assemble into an active protein.

Input1: Pst translocates effector proteins--AvrRpt2

When infecting plants, Pst translocates effector proteins—such as AvrRpt2—into the plant cell cytosol via the Type III Secretion System (T3SS). AvrRpt2 is a putative cysteine protease. Within plant cells, it specifically recognizes the 7-amino acid (7-aa) conserved motif (VPxFGxW) of the host protein RIN4 and cleaves this substrate (Chisholm et al., 2005). The two major cleavage products of RIN4 generated by AvrRpt2 can inhibit PAMPs (pathogen-associated molecular patterns)-triggered immunity (PTI) in plants (Zhao et al., 2021). Meanwhile, the cleavage of RIN4 activates effector-triggered immunity (ETI) mediated by RPS2 (nucleotide-binding site leucine-rich repeats (NB-LRR) protein, resistance to P. syringae 2), triggering the hypersensitive response (HR) and subsequent cell death (Ray et al., 2019).

We aim to design our detection system by leveraging the interaction between AvrRpt2 and peptides from RIN4 proteins, with AvrRpt2 serving as the primary input signal for the system. We convert this signal to generate an output, which is the active recombinant Tobacco Etch Virus protease (TEVp). Herein, we select CC dimerizing domains (segments A, A', and B) with different binding affinities, which were fused to the split protease fragment: the N-terminal fragment of Tobacco Etch Virus protease (nTEVp), the C-terminal fragment of Tobacco Etch Virus protease (cTEVp), and cTEVp* (catalytically inactive mutant of cTEVp), respectively. Additionally, we choose the sequence specifically recognized by AvrRpt2 as the linker between segments A and B. In the presence of AvrRpt2, the A-B linker forms a cleavage site. Subsequently, segments A and A'—which exhibit higher mutual affinity—mediate the reassembly of nTEVp and cTEVp, thereby reconstituting the split TEVp and triggering downstream signaling cascades.

Since the amount of AvrRpt2 injected into plants by Pst is relatively low in the early stage of infection, generating detectable readouts possibly can be challenging. To address this, we need to design a positive feedback loop to amplify the AvrRpt2 signal and lower the limit of detection (LOD)—ultimately enabling more robust output results with a smaller quantity of AvrRpt2. To achieve this, we use TEVp, a protease recognizing a 7-aa specific sequence, to enable signal transduction and amplification (Zheng et al., 2007).

The linker connecting the CC segment A and B —where A is fused to nTEVp and B to cTEVp*—contains not only the AvrRpt2 recognition site but also a TEVp-specific cleavage site (orange-colored in Fig.5). Instead of allowing AvrRpt2 to directly trigger the detection response, we engineered AvrRpt2 to cleave the A-B linker associated with the split TEVp fragments. This cleavage made by AvrRpt2 induces rearrangement between CC modules, thereby reconstituting the catalytic activity of the fused split TEVp. The activated TEVp, in turn, can cleave the orange-colored A-B linker, leading to the generation of more reconstituted active TEVp and amplifying the output signal of the primary input. Concurrently, the activated TEVp cleaves the linker between the C and D segments (as detailed below), which is connected to the N-terminal fragment of Trehalase (nTreA), to initiate downstream responses.

By incorporating this TEVp-mediated positive feedback loop, a small amount of AvrRpt2 can elicit a robust, detectable signal, effectively lowering the LOD of the system.

Input2: phytotoxin--COR

Upon infection of tomato plants by Pst, coronatine (COR), a structural analogue of JA-Ile, is secreted into plant tissues, thereby activating the jasmonic acid (JA) signaling pathway by inducing the binding of Coronatine Insensitive 1 (COI1) and jasmonate ZIM-domain (JAZ) family proteins. Therefore, we used COR as our second input.

The jasmonate signaling pathway is important in plants driven by JA and its derivatives, among which JA-Ile plays an essential role. Under normal conditions, the cytoplasmic level of JA-Ile is low, and the transcription of jasmonate-responsive genes is repressed. However, upon mechanical damage or pathogen infection, JA-Ile accumulates rapidly in the cytoplasm and is transported into the nucleus, where it relieves the repression imposed by the former JAZ complex on related gene expression and promotes the binding of JAZ to COI1 within the SCF complex, initiating subsequent ubiquitination processes of JAZ, which will later be degraded by the 26S-proteasome. (Thines et al., 2007; Wang, Mostafa, Zeng, & Jin, 2021)

Specifically, we focused on the inducible binding process between COI1 and JAZ, and used it to construct our second input pathway, where JA-Ile was replaced by COR. Notably, COR is over 1000 times more effective than JA-Ile at promoting JAZ-COI interaction, making it a good target molecule that is theoretically easy to detect at lower concentrations in the early stages of infection (Katsir, Schilmiller, Staswick, He, & Howe, 2008). It has been verified that the split PPVp fragments fused with CC domains only gained proteolytic activity when an inducer is added, with undetectable leakage in its absence (Fink et al., 2019). Therefore, we expected to replace CC domains with JAZ and COI1 while maintaining a good anti-leakage property. In our design, JAZ1, a member of JAZ proteins from Arabidopsis thaliana, is linked to the C-terminal fragment of plum pox virus protease (cPPVp), while COI1 is connected to the N-terminal fragment of plum pox virus protease (nPPVp). In the presence of COR, COI1 interacts with JAZ1, which in turn facilitates the binding of nPPVp and cPPVp. The resulting complex assumes a complete structure and gains protease activity, thereby activating downstream signaling pathways.

Mechanism of the AND Gate

We designed an "AND gate" logic to ensure that the output system is activated when both the AvrRpt2 signal and the COR signal are present.

Activated by the AvrRpt2 signal, the complete TEVp will cleave the CC-domain linking to the first split protein fragment. Meanwhile, the COR signal promotes the formation of PPVp, which will cut another CC-domain connected to the second split protein fragment. After that, those split protein fragments will self-assemble, leading to the formation of the target signal protein. (Fink et al., 2019)

Output: Selection of Signal Reporters

To achieve diverse visualized output signals, we selected three kinds of split proteins as signal reporters: split luciferase, split Green Fluorescent Protein, and split trehalase.

Split Luciferase

Firstly, we selected luciferase as our output candidate, owing to its high luminescence rate, which can enhance the efficiency of detection.

The intact luciferase is split into an N-terminal fragment (nLuc) and a C-terminal fragment (cLuc). In the initial state, cLuc is bound to chain C, and nLuc is linked to chain C', respectively. Following cleavage by TEVp, which separates chain C and chain D, and PPVp, which separates chain C' and chain E, cLuc and nLuc are released and assemble to form functional luciferase. Then, assembled luciferase can catalyze its luminescent substrate to produce luminescence as an output signal. (Fink et al., 2019)

Split GFP (spGFP)

To reduce costs, we selected spGFP as a secondary output option, as it does not require any exogenous substrate.

In this system, the intact GFP is composed of three split protein fragments: GFP11, GFP10, and GFP1-9. Similar to the split luciferase fragments, GFP11 and GFP10 are linked to chain C and chain C', respectively, while GFP1-9 exists in a free form.

After being cleaved by TEVp and PPVp, GFP10 and GFP11 fragments are released. Then they will spontaneously assemble with GFP1-9 to form functional GFP, which is capable of emitting fluorescent signals. (Fink et al., 2019)

Split Trehalase

The detection of those previous two output candidates relies on fluorescence, which is susceptible to interference from light, temperature, pH, and other environmental factors. This poses significant practical challenges for on-farm applications: outdoor conditions such as sunlight and rainfall can easily distort fluorescent signals, while the equipment required for fluorescence detection, paired with the need for manual light observation, may also be cumbersome for users in practical production.

In real-world agricultural production, farmers prefer simple and stable visual signals—ones that should be intuitive, requiring no additional devices or extra operations. Therefore, taking this specific demand of growers into account, we decided to use glucose as a visualized output signal. Compared with light signals, glucose is not only easier to detect using common tools like glucose meters, but also resistant to interference from environmental factors. This simplifies operation for farmers: instead of identifying faint light signals, they only need to check the glucose indicators we provide.

This consideration led us to our final selection: split trehalse. Similar to luciferase, we use split trehalase as the enzyme and trehalose as the substrate. The intact trehalase (TreA) is split into N-terminal fragment (nTreA) and C-terminal fragment (cTreA). After the cleavage of TEVp and PPVp, cTreA and nTreA will form functional trehalase, which can catalyze trehalose into glucose. After that, all we need to do is detect the glucose content, whose signal reading is more convenient and accessible. (Drikic & De Buck, 2018)

References

1. Reinke, A. W., Grant, R. A., & Keating, A. E. (2010). A synthetic Coiled-Coil interactome provides heterospecific modules for molecular engineering. Journal of the American Chemical Society, 132(17), 6025-6031.
2. Oakley, M. (2001). The design of antiparallel coiled coils. Current Opinion in Structural Biology, 11(4), 450-457.
3. Chisholm, S. T., Dahlbeck, D., Krishnamurthy, N., Day, B., Sjolander, K., & Staskawicz, B. J. (2005). Molecular characterization of proteolytic cleavage sites of the Pseudomonas syringae effector AvrRpt2. Proceedings of the National Academy of Sciences, 102(6), 2087-2092.
4. Zhao, G., Guo, D., Wang, L., Li, H., Wang, C., & Guo, X. (2021). Functions of RPM1-interacting protein 4 in plant immunity. Planta, 253(1).
5. Ray, S. K., Macoy, D. M., Kim, W., Lee, S. Y., & Kim, M. G. (2019). Role of RIN4 in Regulating PAMP-Triggered Immunity and Effector-Triggered Immunity: Current Status and Future Perspectives. PubMed, 42(7), 503-511.
6. Zheng, N., De Jesús Pérez, J., Zhang, Z., Domínguez, E., Garcia, J. A., & Xie, Q. (2007). Specific and efficient cleavage of fusion proteins by recombinant plum pox virus NIa protease. Protein Expression and Purification, 57(2), 153-162.
7. Thines, B., Katsir, L., Melotto, M., Niu, Y., Mandaokar, A., Liu, G., . . . Browse, J. (2007). JAZ repressor proteins are targets of the SCF(COI1) complex during jasmonate signalling. Nature, 448(7154), 661-665.
8. Wang, Y., Mostafa, S., Zeng, W., & Jin, B. (2021). Function and Mechanism of Jasmonic Acid in Plant Responses to Abiotic and Biotic Stresses. Int J Mol Sci, 22(16).
9. Katsir, L., Schilmiller, A. L., Staswick, P. E., He, S. Y., & Howe, G. A. (2008). COI1 is a critical component of a receptor for jasmonate and the bacterial virulence factor coronatine. Proceedings of the National Academy of Sciences, 105(19), 7100-7105.
10. Fink, T., Lonzarić, J., Praznik, A., Plaper, T., Merljak, E., Leben, K., . . . Jerala, R. (2019). Design of fast proteolysis-based signaling and logic circuits in mammalian cells. Nat Chem Biol, 15(2), 115-122.
11. Drikic, M., & De Buck, J. (2018). Split trehalase as a versatile reporter for a wide range of biological analytes. Biotechnology and Bioengineering, 115(5), 1128-1136.